1 00:00:02,670 --> 00:00:09,390 In this tutorial, we will be creating an application where I will be counting the persons entering 2 00:00:09,390 --> 00:00:11,610 and leaving from a specific area. 3 00:00:11,610 --> 00:00:18,810 Plus, I will also show you how you can count the number of vehicles entering and leaving from a specific 4 00:00:18,810 --> 00:00:20,220 area as well. 5 00:00:20,730 --> 00:00:27,750 So in this tutorial we will be doing objects counting and an object will be counting the objects entering 6 00:00:27,750 --> 00:00:29,490 and leaving from a specific area. 7 00:00:29,490 --> 00:00:32,430 And we will be creating two different applications. 8 00:00:32,430 --> 00:00:39,210 One is the upper one is the count of persons entering and leaving from a specific area, and the second 9 00:00:39,210 --> 00:00:44,610 one is the count of vehicles which are entering and leaving from a specific area. 10 00:00:44,610 --> 00:00:46,050 So let's get started with it. 11 00:00:46,050 --> 00:00:52,800 So in this tutorial, uh, with these applications I will be using Yolov5 and Deepsort object tracking 12 00:00:52,800 --> 00:00:53,490 algorithm. 13 00:00:53,790 --> 00:00:59,010 So you will be in a state of the art object detection algorithm, while, uh, Deepsort is the state 14 00:00:59,010 --> 00:01:01,980 of the art object tracking algorithm. 15 00:01:01,980 --> 00:01:02,190 Okay. 16 00:01:03,910 --> 00:01:10,150 So we will be integrating object detection and object tracking so that we can create a card counting 17 00:01:10,150 --> 00:01:13,030 application or vehicle counting application. 18 00:01:13,330 --> 00:01:19,510 Or you can say we will be, uh, counting the persons entering and leaving from a specific area. 19 00:01:19,510 --> 00:01:20,920 So there will be two applications. 20 00:01:20,920 --> 00:01:26,860 One is the persons entering and leaving, or the count of persons entering and leaving from a specific 21 00:01:26,860 --> 00:01:27,160 area. 22 00:01:27,160 --> 00:01:32,560 And the second one will be the count of vehicles entering and leaving from a specific area. 23 00:01:34,060 --> 00:01:38,260 So first of all, before we go ahead, uh, we have lost. 24 00:01:40,280 --> 00:01:42,470 So there are so. 25 00:01:43,560 --> 00:01:46,230 So before, uh, let's get started with it. 26 00:01:46,230 --> 00:01:51,930 So first of all, you can see over here I have the YOLO v nine repository or over here. 27 00:01:51,930 --> 00:01:55,140 So I will just first clone this YOLO nine GitHub repo. 28 00:01:55,140 --> 00:01:58,110 So I will just open the PyCharm from here. 29 00:01:58,950 --> 00:02:01,200 So you can see here the PyCharm ID. 30 00:02:01,230 --> 00:02:06,120 I've just created a new project by the name entering leaving counting your Deepsort recording. 31 00:02:06,120 --> 00:02:10,320 So you I've just created this project in my F local directory. 32 00:02:10,650 --> 00:02:11,220 Okay. 33 00:02:11,220 --> 00:02:13,230 So let's get started with it. 34 00:02:14,510 --> 00:02:18,650 So first of all, I just, uh, I will just note that you have the benign GitHub repository. 35 00:02:18,650 --> 00:02:22,070 So I will just add the link over here and write git clone. 36 00:02:24,250 --> 00:02:28,240 So now you can see that, uh, your login and GitHub repository has been cloned. 37 00:02:28,630 --> 00:02:29,110 Okay. 38 00:02:29,110 --> 00:02:32,290 So if I just refresh this over here. 39 00:02:34,170 --> 00:02:34,470 It. 40 00:02:37,760 --> 00:02:38,750 But it's not here. 41 00:02:39,110 --> 00:02:39,350 Okay. 42 00:02:39,350 --> 00:02:44,420 So you can see here we have the lone Euro v nine, uh, complete GitHub repository over here. 43 00:02:44,420 --> 00:02:49,820 So you can see all the files that we have in the repository are being here. 44 00:02:49,820 --> 00:02:55,580 So now I will just set this yolo v nine uh directory uh folder as my parent directory. 45 00:02:55,580 --> 00:02:58,910 So to do this I will just write CD which stands for current directory. 46 00:02:58,910 --> 00:03:02,870 And I will just set this folder as my current directory. 47 00:03:02,870 --> 00:03:04,130 And now it's being set. 48 00:03:04,130 --> 00:03:09,770 So before we go ahead let's create our virtual environment so you can create the virtual environment 49 00:03:09,770 --> 00:03:13,340 if you don't want to disturb your uh packages. 50 00:03:13,340 --> 00:03:16,580 So you will just write conda create minus n. 51 00:03:16,580 --> 00:03:20,270 And here I will just write the name of my virtual environment as v nine. 52 00:03:21,110 --> 00:03:21,620 Okay. 53 00:03:23,830 --> 00:03:27,490 So I am just creating a virtual environment by the name your Logging line. 54 00:03:28,240 --> 00:03:29,260 Okay, so. 55 00:03:32,500 --> 00:03:37,030 So you can see that it says that already a virtual environment by the name YOLO v9. 56 00:03:37,030 --> 00:03:39,670 So I've revoked that already created environment. 57 00:03:39,670 --> 00:03:41,710 And now I'm just waiting a new environment. 58 00:03:41,710 --> 00:03:48,250 So by the name of ignite and I will just write conda activate yolo v9. 59 00:03:48,670 --> 00:03:49,540 So that's it. 60 00:03:50,590 --> 00:03:53,800 So now I have just created a virtual world by the name yolo v9. 61 00:03:53,800 --> 00:04:01,420 And now I will install all the packages that are mentioned in the requirements.txt file. 62 00:04:01,420 --> 00:04:04,060 So over here you can see the requirements.txt file. 63 00:04:04,060 --> 00:04:10,060 In the requirements.txt file you will find all the packages uh, that are required to run this uh to 64 00:04:10,060 --> 00:04:12,940 run object detection object tracking scripts successfully. 65 00:04:15,150 --> 00:04:17,220 So if I just write pip install minus r. 66 00:04:17,220 --> 00:04:21,570 So this will install all the packages that are listed inside this requirements.txt file. 67 00:04:21,570 --> 00:04:25,260 So if you install these packages one by one this will take quite some time. 68 00:04:25,260 --> 00:04:30,510 So it's always a good practice to list all your packages in the requirements.txt file and books. 69 00:04:30,780 --> 00:04:35,160 And by writing one command in the terminal pip install minus r requirements.txt. 70 00:04:35,220 --> 00:04:40,500 So this install all the packages that are listed in the requirements dot txt file. 71 00:04:40,980 --> 00:04:43,890 So the package installation will take around 3 to 4 minutes. 72 00:04:43,890 --> 00:04:46,920 So let's get back as this packages gets drawn. 73 00:04:51,170 --> 00:04:54,320 Now you can see that all the packages are being installed. 74 00:04:54,320 --> 00:04:59,180 So the instructions are I've just created a notepad file by the name instruction. 75 00:04:59,510 --> 00:05:03,410 Here I have listed all the steps that we will be following in this tutorial. 76 00:05:03,410 --> 00:05:06,860 So now we have uploaded the nine GitHub repo. 77 00:05:06,860 --> 00:05:08,630 We have created a virtual environment. 78 00:05:08,630 --> 00:05:10,820 We have installed all the required packages. 79 00:05:10,820 --> 00:05:12,410 So three things are done. 80 00:05:12,410 --> 00:05:16,310 So now we will download the sample videos from the Google Drive link below. 81 00:05:16,310 --> 00:05:23,090 So I have added some uh videos or to test uh when we try to. 82 00:05:23,090 --> 00:05:29,990 So one is the person, uh, walking on a street and other are the vehicles on a highway lane so that 83 00:05:29,990 --> 00:05:35,000 we can, uh, count the vehicles entering and leaving and count the persons entering and leaving the 84 00:05:35,000 --> 00:05:35,900 specific area. 85 00:05:35,900 --> 00:05:39,140 So I have placed those demo videos onto my Google Drive. 86 00:05:39,140 --> 00:05:44,480 So I will directly download those demo videos from Google Drive into this PyCharm, uh, IDE. 87 00:05:44,690 --> 00:05:50,300 So to download the demo videos from, uh, Google Drive and directly into the PyCharm IDE, we need 88 00:05:50,300 --> 00:05:53,180 to install the pip install g down package. 89 00:05:53,450 --> 00:05:55,940 So I will just write over here pip install a down. 90 00:05:56,180 --> 00:06:01,700 And I will add this package in the uh requirements.txt file over here as well. 91 00:06:01,700 --> 00:06:02,360 So. 92 00:06:04,790 --> 00:06:06,980 Now you can see this package is installed. 93 00:06:06,980 --> 00:06:15,740 So this package now allows us to download any demo videos or any files or any image file or any video 94 00:06:15,740 --> 00:06:23,150 and image file or any, uh, other files, zip files from the Google Drive directly into our local IDE. 95 00:06:23,180 --> 00:06:28,550 It can be Visual Studio Video, it can be PyCharm, or it can be Orange Studio or any other IDE. 96 00:06:28,580 --> 00:06:35,390 Okay, so I've just added the write down and edit the link over here. 97 00:06:35,390 --> 00:06:42,890 So now you can see that uh, just watch that mp4 uh is downloaded into our PyCharm IDE. 98 00:06:43,100 --> 00:06:47,510 Or if I just refresh this up you can see the text box dot mp4 file over here. 99 00:06:47,510 --> 00:06:49,580 And let me just open this up. 100 00:06:49,580 --> 00:06:53,450 So now I will download the second demo video from the Google Drive. 101 00:06:53,630 --> 00:06:56,060 Like look directly into this PyCharm ID. 102 00:06:56,060 --> 00:06:58,340 So I will just add this link over here. 103 00:07:01,370 --> 00:07:04,400 So now you can see the second demo video is being downloaded. 104 00:07:04,790 --> 00:07:08,240 Uh, the name of this demo video is S3 dot mp4. 105 00:07:08,780 --> 00:07:13,970 Okay, so we, I have downloaded two demo videos and that would be enough to test our scripts. 106 00:07:14,360 --> 00:07:18,110 Okay so we have test for MP4 and testing. 107 00:07:22,310 --> 00:07:22,640 Okay. 108 00:07:22,640 --> 00:07:23,600 So here we have a test. 109 00:07:23,600 --> 00:07:24,080 We know that. 110 00:07:24,980 --> 00:07:26,390 So that's good enough. 111 00:07:26,390 --> 00:07:32,810 So now if we just see over here uh, I have cloned a GitHub repo, created a virtual environment, installed 112 00:07:32,810 --> 00:07:36,230 all the required packages, download the sample video. 113 00:07:36,230 --> 00:07:41,420 So I've downloaded two sample videos from the Google Drive into my PyCharm IDE as well. 114 00:07:41,420 --> 00:07:47,480 So now in next step, in the first step I will be doing object tracking using YOLO, v9 and Sword algorithm. 115 00:07:47,480 --> 00:07:49,190 So let's get started with it. 116 00:07:53,770 --> 00:07:54,220 Up here. 117 00:07:54,220 --> 00:07:58,540 You can see I've just created a file by the name detect-dual-tracking.py. 118 00:07:58,840 --> 00:08:02,590 Currently it only has the object detection board in the first step. 119 00:08:02,740 --> 00:08:08,350 I will integrate a Deepsort object tracking with yolo v9 object detection board over here. 120 00:08:08,350 --> 00:08:11,230 So currently this is all object detection code. 121 00:08:11,260 --> 00:08:16,810 We will first integrate object tracking and then we will be counting the person or vehicles entering 122 00:08:16,810 --> 00:08:17,980 and leaving in the next step. 123 00:08:17,980 --> 00:08:23,590 So object tracking using YOLO v9 and the source. 124 00:08:23,620 --> 00:08:25,420 So here right. 125 00:08:25,720 --> 00:08:30,340 So we will be doing object tracking using yolo v9 and Deepsort algorithm. 126 00:08:30,490 --> 00:08:35,200 So let me integrate object tracking board into this object detection board over here. 127 00:08:38,930 --> 00:08:41,870 The midfielders open the stands over here. 128 00:08:41,870 --> 00:08:44,360 So from the algorithm. 129 00:08:44,750 --> 00:08:50,030 The algorithm gives us output in the form of bounding box coordinates x one and y one represents the 130 00:08:50,570 --> 00:08:53,870 top, left corner and bounding box coordinates, and x two. 131 00:08:53,870 --> 00:08:57,920 Y two represents the bottom right corner bounding box coordinates. 132 00:08:57,920 --> 00:09:02,330 So this is the output uh, which we get from the algorithm. 133 00:09:02,330 --> 00:09:05,660 Plus we also get the confidence score and the class name okay. 134 00:09:05,660 --> 00:09:13,130 So but uh, we will be passing the output that we get from the algorithm as an input to the Deepsort 135 00:09:13,130 --> 00:09:13,640 algorithm. 136 00:09:13,640 --> 00:09:19,400 So the output that we get from the YOLO algorithm will be passed as an input to the Deepsort object 137 00:09:19,400 --> 00:09:20,360 tracking algorithm. 138 00:09:20,360 --> 00:09:26,090 So the Deepsort object tracking algorithm accepts uh, the input as height. 139 00:09:26,630 --> 00:09:27,650 And so. 140 00:09:28,420 --> 00:09:34,240 The Deepsort algorithm does not accept the bounding box coordinates like x one, y one or x two, y 141 00:09:34,270 --> 00:09:34,510 two. 142 00:09:34,540 --> 00:09:40,990 The Deepsort object tracking algorithm accepts input as the height and width of the bounding box. 143 00:09:40,990 --> 00:09:41,500 Okay. 144 00:09:41,530 --> 00:09:44,590 Last, the center point of the bounding box. 145 00:09:44,590 --> 00:09:49,090 So this is what our Deepsort object tracking algorithm accept as input. 146 00:09:49,330 --> 00:09:52,300 Along with this, accept a confidence score and class name. 147 00:09:52,300 --> 00:09:58,210 So the confidence score and class name are same uh, for the algorithm. 148 00:09:58,210 --> 00:10:02,920 But uh, the Yolo v nine algorithm gives us output in the form of bounding box coordinates. 149 00:10:02,920 --> 00:10:09,970 So Deepsort algorithm will not accept the, uh, output input uh, in the form of bounding box coordinates. 150 00:10:09,970 --> 00:10:14,530 So as we pass the yolo v nine output as an input to that Deepsort algorithm. 151 00:10:14,530 --> 00:10:22,120 So we need to convert this, uh, bounding box coordinates x one, y one and X2Y2 into this form like 152 00:10:22,120 --> 00:10:25,900 height and width of the bounding box and the center point of the bounding box. 153 00:10:25,900 --> 00:10:32,680 So before it pass the YOLO v nine output as an input to the Deepsort algorithm, we need to transform 154 00:10:32,680 --> 00:10:38,290 the Yolo nine output into the format that is required by the Deepsort algorithm, which is height and 155 00:10:38,290 --> 00:10:41,350 width of the bounding box and center coordinates of the bounding box. 156 00:10:41,710 --> 00:10:42,730 So let's do it. 157 00:10:58,620 --> 00:10:59,640 So x. 158 00:10:59,820 --> 00:11:02,520 So this is X1Y1 and X2Y2. 159 00:11:02,910 --> 00:11:03,480 Okay. 160 00:11:04,920 --> 00:11:07,170 So last we have the class names. 161 00:11:09,040 --> 00:11:09,280 Okay. 162 00:11:12,230 --> 00:11:14,690 But this is what we get from the revolving line output. 163 00:11:14,720 --> 00:11:17,450 So the output is currently in the form of tensors. 164 00:11:17,450 --> 00:11:20,630 We need to convert the output from tensors into integers. 165 00:11:55,350 --> 00:12:01,230 The next, we will find the center coordinates for each of the bounding box. 166 00:12:18,080 --> 00:12:22,190 So you can see over here we have find the center coordinates for each of the bounding box. 167 00:12:22,190 --> 00:12:27,770 As I told you, uh, that Deepsort algorithm accepts input as a center coordinates of the bounding box 168 00:12:27,770 --> 00:12:29,600 and height and width of the bounding box. 169 00:12:29,600 --> 00:12:32,510 So next we need to find the height and width of the bounding box. 170 00:12:33,140 --> 00:12:34,760 So I will just write. 171 00:12:53,090 --> 00:12:53,300 The. 172 00:12:58,730 --> 00:13:02,840 So here you can see that we have pointed out the width and the height of the bounding boxes. 173 00:13:02,960 --> 00:13:10,400 So, uh, if you just see over here so you can see here, uh, this is the x one minus x two. 174 00:13:10,520 --> 00:13:12,590 We do this, we get the width. 175 00:13:12,590 --> 00:13:15,110 And if we do y one minus y two we get the height. 176 00:13:15,440 --> 00:13:18,080 So I'm just following the same steps over here. 177 00:13:21,720 --> 00:13:26,910 So now, uh, as in a single frame, there will be multiple different objects will be detected. 178 00:13:26,910 --> 00:13:32,910 So now we will need to append this width and height and center coordinates of the bounding box for a 179 00:13:32,910 --> 00:13:34,440 complete frame into our list. 180 00:13:34,440 --> 00:13:35,760 So we will do this. 181 00:13:43,950 --> 00:13:46,470 I would create three empty lists over here. 182 00:13:55,560 --> 00:13:57,690 So this is what the bounding box coordinates. 183 00:13:57,690 --> 00:13:58,980 This is for the confidence score. 184 00:13:59,010 --> 00:14:02,280 This is for the class name and one we have for output as well. 185 00:14:17,960 --> 00:14:18,230 The. 186 00:14:24,210 --> 00:14:30,810 Got a Deepsort algorithm accept input as in the form of central coordinates Cxxvi bounding box coordinates. 187 00:14:30,810 --> 00:14:32,640 Width and height. 188 00:14:35,840 --> 00:14:41,270 Well, now I will just append these values to this list which we have created. 189 00:14:44,400 --> 00:14:47,160 Now I will also find out the quantum score for this. 190 00:14:47,160 --> 00:14:51,930 So I will just use for the math library so that I can convert the confidence score from tensors into 191 00:14:51,930 --> 00:14:52,380 integer. 192 00:14:56,010 --> 00:14:56,280 He. 193 00:15:02,560 --> 00:15:06,580 And now I will also open the scoring to the list which I have created above. 194 00:15:14,520 --> 00:15:18,060 And I'm just converting the class IDs into integers as well. 195 00:15:18,090 --> 00:15:21,900 Because they are originally in the form of tensors, we need to convert them into integers. 196 00:15:24,480 --> 00:15:28,110 And now I'm just spilling all the class ideas into the list as well. 197 00:15:34,090 --> 00:15:35,230 But now I'm just converted. 198 00:15:35,440 --> 00:15:42,400 Uh, so now I've just, uh, transformed my Yolo V8 output into the format that is required by the Deepsort 199 00:15:42,400 --> 00:15:43,000 algorithm. 200 00:15:43,030 --> 00:15:48,670 Like, you can see that I have the width and height and center coordinates of the bounding boxes. 201 00:15:48,670 --> 00:15:50,020 I have the confidence score. 202 00:15:50,080 --> 00:15:52,450 I have the class names as well. 203 00:15:52,450 --> 00:15:56,890 So now we need to convert, uh, this output into tensors. 204 00:15:56,980 --> 00:16:01,210 Uh, or this, uh, this, uh, output into dancers. 205 00:16:01,210 --> 00:16:04,780 And then we'll pass in this as an input to the Deepsort algorithm. 206 00:16:41,060 --> 00:16:41,300 Okay. 207 00:16:41,300 --> 00:16:44,180 So we have converted the this as. 208 00:16:45,140 --> 00:16:47,240 So we have going to just enter coordinates. 209 00:16:47,270 --> 00:16:51,080 Uh bounding boxes width and height into tensors that are required. 210 00:16:51,110 --> 00:16:55,700 We don't require to convert the class IDs into tensors currently. 211 00:16:57,380 --> 00:17:00,920 So I will just write output is equal to okay. 212 00:17:01,040 --> 00:17:03,350 So okay okay okay. 213 00:17:03,350 --> 00:17:07,790 So now I have not initialized with over here. 214 00:17:07,790 --> 00:17:10,400 So before we go ahead here. 215 00:17:16,420 --> 00:17:16,960 Over here. 216 00:17:16,960 --> 00:17:20,050 I've just created a file by the name object counter dot pi. 217 00:17:20,380 --> 00:17:23,560 So here you can see that I have class object counter. 218 00:17:23,560 --> 00:17:27,700 And over here I will explain this complete code which is over here. 219 00:17:27,700 --> 00:17:30,310 So first we have this initialize Deepsort. 220 00:17:30,310 --> 00:17:34,810 So first I will do what I will do here is okay. 221 00:17:34,810 --> 00:17:39,310 So I will just write from check out the import. 222 00:17:39,820 --> 00:17:42,250 So I will just import this class. 223 00:17:42,250 --> 00:17:47,260 So all the functions that we are required are inside this class okay. 224 00:17:47,830 --> 00:17:50,980 So I will just write object counter. 225 00:17:51,100 --> 00:17:53,920 Now I will just need to the initialization over here. 226 00:18:04,110 --> 00:18:04,530 Okay. 227 00:18:04,650 --> 00:18:05,760 We are good to go. 228 00:18:05,880 --> 00:18:06,720 Next I will write. 229 00:18:06,750 --> 00:18:08,490 Each sword is equal to. 230 00:18:11,170 --> 00:18:15,400 Object counter, and we will be using the initialized Deepsort function over here. 231 00:18:15,790 --> 00:18:16,390 Okay. 232 00:18:21,540 --> 00:18:24,720 So now what I will do is. 233 00:18:26,120 --> 00:18:29,330 Oh, well, I will just go over here. 234 00:18:30,610 --> 00:18:31,030 It's not. 235 00:18:31,420 --> 00:18:33,340 So if I just show you. 236 00:18:33,370 --> 00:18:37,420 Okay, so now I have don't have added the Deepsort files over here as well. 237 00:18:37,420 --> 00:18:39,640 So let me add those Deepsort files. 238 00:18:40,360 --> 00:18:43,150 And so I haven't added the link over here. 239 00:18:43,150 --> 00:18:43,690 So. 240 00:18:46,520 --> 00:18:50,780 Before we write the sixth and seventh step, we need to add. 241 00:18:51,980 --> 00:18:53,690 We need to download. 242 00:18:54,880 --> 00:18:58,270 That Deepsort files are. 243 00:19:00,370 --> 00:19:01,210 Right. 244 00:19:02,530 --> 00:19:02,950 It. 245 00:19:08,300 --> 00:19:09,170 We can see here. 246 00:19:09,170 --> 00:19:10,910 I did a Google Drive link. 247 00:19:10,910 --> 00:19:16,250 I have pasted each zip file in this Google Drive link on this Google Drive link, so I will actually 248 00:19:16,250 --> 00:19:21,620 download a zip file from this Google Drive link into my PyCharm ID, and then I will unzip this file. 249 00:19:22,340 --> 00:19:22,820 Okay. 250 00:19:26,100 --> 00:19:30,990 So I'm downloading this deepsort zip file from the Google Drive directly into this PyCharm IDE. 251 00:19:30,990 --> 00:19:35,340 And uh, now you can see the file size is 43.1 MB. 252 00:19:35,340 --> 00:19:38,640 And as I as it gets downloaded I will unzip this file. 253 00:19:42,030 --> 00:19:42,300 Okay. 254 00:19:42,300 --> 00:19:44,580 So now you can see the file is being downloaded. 255 00:19:44,610 --> 00:19:47,250 Let me just unzip it. 256 00:19:50,550 --> 00:19:52,200 So now unzip that file. 257 00:19:52,200 --> 00:19:55,650 So let's go ahead and you can see I have this file over here. 258 00:19:55,650 --> 00:20:01,140 So if I just open the deepsort uh.py code over here. 259 00:20:01,140 --> 00:20:04,140 So now you can see here we have the class deepsort. 260 00:20:04,140 --> 00:20:06,720 And here we have the initialize function over here. 261 00:20:06,720 --> 00:20:10,500 And I will be using this Deepsort dot update function. 262 00:20:10,500 --> 00:20:12,750 And here I am this update function. 263 00:20:12,750 --> 00:20:17,670 Now you can see that we have the input in the form of bounding box coordinates and width. 264 00:20:17,670 --> 00:20:23,520 And uh, we have the input in the form of width and the height of the bounding box and center coordinates 265 00:20:23,520 --> 00:20:24,960 of the bounding boxes. 266 00:20:24,960 --> 00:20:31,500 Plus we have the confidence score and the class name and the original frame as the input of this update 267 00:20:31,500 --> 00:20:31,980 function. 268 00:20:31,980 --> 00:20:32,730 So that's good. 269 00:20:35,010 --> 00:20:40,950 So I will be using this Deepsort dot update function and in the import over here. 270 00:20:43,570 --> 00:20:44,980 In the input over here. 271 00:20:44,980 --> 00:20:49,600 I will be passing this center coordinates of the bounding box bounding box width and the bounding box 272 00:20:49,600 --> 00:20:49,930 height. 273 00:20:49,930 --> 00:20:52,630 And we have all this inside this. 274 00:20:55,860 --> 00:21:02,550 Inside this because we have converted the this into tensors, and now we are passing this as an input 275 00:21:02,550 --> 00:21:05,190 to the data sort update function. 276 00:21:05,790 --> 00:21:06,510 Okay. 277 00:21:06,690 --> 00:21:08,250 For the input ready sort algorithm. 278 00:21:08,250 --> 00:21:14,670 Plus we will be passing the confidence port and the class IDs. 279 00:21:22,130 --> 00:21:26,090 And we will be passing the grand frame which I have already shown. 280 00:21:26,630 --> 00:21:33,290 Okay, so this will be the input to the deep sort algorithm, and we will be passing this to that update 281 00:21:33,290 --> 00:21:35,090 function that I've already shown you. 282 00:21:40,790 --> 00:21:45,860 So next what we'll do is we have the length of the outputs greater than zero. 283 00:21:45,860 --> 00:21:52,100 Like if there are some objects, if any of the object is directly in the current frame, like 1 or 2, 284 00:21:53,270 --> 00:21:54,440 then what we. 285 00:21:54,440 --> 00:21:54,800 Right. 286 00:21:54,800 --> 00:22:01,010 So I mean, so if you just print out the output from here. 287 00:22:09,270 --> 00:22:12,030 Let me show you how our output looks like. 288 00:22:21,370 --> 00:22:25,360 Thought we were just running this detect as dual tracking, not by script. 289 00:22:25,780 --> 00:22:34,420 And that we just got this text box, not MP4 video, which we have downloaded already from the Google 290 00:22:34,420 --> 00:22:34,900 Drive link. 291 00:22:34,900 --> 00:22:38,230 So let me show you. 292 00:22:38,230 --> 00:22:41,320 So if I just show you here. 293 00:22:43,370 --> 00:22:49,310 From the Deepsort algorithm like we have passed it in the in as a center coordinates of the bounding 294 00:22:49,310 --> 00:22:55,790 boxes and width and the height of the bounding boxes as an input to the dot algorithm and from the from 295 00:22:55,790 --> 00:22:57,650 the output of the algorithm. 296 00:22:57,830 --> 00:22:59,660 What we get is the bounding box coordinates. 297 00:22:59,660 --> 00:23:02,540 So this is the output of the Deepsort algorithm. 298 00:23:02,540 --> 00:23:05,480 We get the bounding box coordinates a unique ID. 299 00:23:05,480 --> 00:23:09,560 Basically, in object tracking we assign a unique ID to each of the detected object. 300 00:23:09,560 --> 00:23:12,920 So we get the unique ID for each of the detected object. 301 00:23:12,920 --> 00:23:16,250 Plus we get the class name or class IDs. 302 00:23:16,250 --> 00:23:18,860 For example, for the person class, the class ID is zero. 303 00:23:18,860 --> 00:23:24,170 So the Deepsort algorithm returns us the bounding box coordinates like x one, y one, and x two, y 304 00:23:24,170 --> 00:23:29,300 two that are the top left corner bounding box coordinate and bottom right corner bounding box coordinates. 305 00:23:29,300 --> 00:23:32,810 Along with this Deepsort algorithm, output gives us signs. 306 00:23:32,810 --> 00:23:37,760 In the Deepsort algorithm output, we get a unique ID for each of the detected object. 307 00:23:37,760 --> 00:23:39,530 Basically, this is what object tracking do. 308 00:23:39,530 --> 00:23:42,950 In object tracking, we assign a unique ID to each detected object. 309 00:23:42,950 --> 00:23:44,570 Thus we get the class ID. 310 00:23:44,570 --> 00:23:47,420 For example, for the person class, the class ID is zero. 311 00:23:47,690 --> 00:23:50,900 So this is the output that we get from the Deepsort algorithm. 312 00:23:53,660 --> 00:23:54,590 Love it. 313 00:24:01,290 --> 00:24:04,590 So I have got the idea for this error because. 314 00:24:06,600 --> 00:24:10,260 So we just layer this over here. 315 00:24:13,190 --> 00:24:13,490 It. 316 00:24:16,000 --> 00:24:16,810 The. 317 00:24:17,890 --> 00:24:19,300 Okay, so now it's working. 318 00:24:19,300 --> 00:24:19,810 Okay. 319 00:24:20,230 --> 00:24:23,350 So now I think. 320 00:24:29,950 --> 00:24:30,370 Yeah. 321 00:24:30,370 --> 00:24:31,810 Hopefully it will work fine. 322 00:24:36,530 --> 00:24:38,240 Let's see how it goes. 323 00:24:39,170 --> 00:24:41,330 I think now the issue will be solved. 324 00:24:42,470 --> 00:24:43,850 Okay, so. 325 00:24:51,370 --> 00:24:55,510 Okay, so we also need to run the Deepsort algorithm step. 326 00:24:55,510 --> 00:24:59,620 We need to have the easy dish package installed as well. 327 00:24:59,620 --> 00:25:01,690 So I try to. 328 00:25:04,150 --> 00:25:08,770 Then I can just, like, maybe start easy then. 329 00:25:09,640 --> 00:25:09,880 Yeah. 330 00:25:10,840 --> 00:25:12,220 So I run over here as well. 331 00:25:12,220 --> 00:25:14,440 So if you just install all these packages. 332 00:25:14,770 --> 00:25:20,380 Uh, if you run install the requirements.txt file, this package will also get installed along the way. 333 00:25:21,490 --> 00:25:22,780 So let's run this up. 334 00:25:28,900 --> 00:25:34,090 So this will take a few seconds because on my local system I don't have GPU, I have CPU, so I'm running 335 00:25:34,090 --> 00:25:34,810 on CPU. 336 00:25:44,800 --> 00:25:44,980 A. 337 00:25:48,630 --> 00:25:52,590 Or last one will have the output being shown over here. 338 00:25:58,880 --> 00:25:59,210 Okay. 339 00:25:59,210 --> 00:26:02,120 So last data packet is shown. 340 00:26:02,120 --> 00:26:08,870 Like let me just correct this up over here as well where I have num pi. 341 00:26:12,680 --> 00:26:14,030 It's all we need to have. 342 00:26:14,030 --> 00:26:14,810 NumPy. 343 00:26:15,320 --> 00:26:17,570 Less than 1.24. 344 00:26:17,600 --> 00:26:19,820 Else will give an error log. 345 00:26:19,880 --> 00:26:24,290 Okay, so we will I will first uninstall numpy. 346 00:26:28,740 --> 00:26:31,590 Because with the latest version of numpy we get error. 347 00:26:31,590 --> 00:26:33,900 So we'll just downgrade the one version. 348 00:26:34,980 --> 00:26:35,790 So. 349 00:26:40,050 --> 00:26:43,980 Or you just install numpy with version less than. 350 00:26:45,800 --> 00:26:48,650 1.2 score. 351 00:26:49,550 --> 00:26:52,160 Okay, so I hope this will fix up the issue. 352 00:27:39,120 --> 00:27:42,480 Currently there is no output was. 353 00:27:42,480 --> 00:27:45,810 As we go out in the next frame we will be able to see the output. 354 00:27:45,810 --> 00:27:48,570 I am not showing displaying the output currently. 355 00:27:49,230 --> 00:27:50,880 Okay, so now you can see here. 356 00:27:50,880 --> 00:27:59,550 Uh, so this is the output that we get from the uh sort algorithm like you can see over here. 357 00:28:03,750 --> 00:28:07,770 So this is the output that we are getting from the Deepsort algorithm, so that these are the bounding 358 00:28:07,770 --> 00:28:08,850 box coordinates. 359 00:28:09,270 --> 00:28:09,930 Okay. 360 00:28:10,260 --> 00:28:14,760 Here you can see this is a unique ID that is being assigned to each of the detected object. 361 00:28:14,760 --> 00:28:16,920 And here we have the class IDs. 362 00:28:16,920 --> 00:28:19,380 For example for the person class the class ID zero. 363 00:28:19,380 --> 00:28:23,250 So these are the bounding box coordinates X1Y1X2Y2. 364 00:28:23,490 --> 00:28:28,200 This is the unique ID like in object tracking we assign a unique ID to each of the detected object. 365 00:28:28,200 --> 00:28:28,740 So the. 366 00:28:28,740 --> 00:28:30,480 Here the unique id is one. 367 00:28:30,480 --> 00:28:32,820 Here for the second object the unique id is two. 368 00:28:32,880 --> 00:28:34,650 For the third object, unique id is four. 369 00:28:34,650 --> 00:28:38,670 And this is the class ID for the person class. 370 00:28:38,670 --> 00:28:40,410 The class ID is zero. 371 00:28:40,410 --> 00:28:40,800 Okay. 372 00:28:40,800 --> 00:28:43,020 So if I just show you over here. 373 00:28:44,070 --> 00:28:48,600 So the Deepsort algorithm gives us the bounding box coordinates, a unique ID for each of the detected 374 00:28:48,600 --> 00:28:48,990 object. 375 00:28:48,990 --> 00:28:50,640 Plus we get the class IDs. 376 00:28:50,880 --> 00:28:54,570 For example, for the person class the class id zero okay. 377 00:28:54,570 --> 00:28:57,330 So this is for what we are getting for each of the frame. 378 00:28:57,330 --> 00:29:00,720 Like the complete video is divided into 341 frames. 379 00:29:00,720 --> 00:29:03,690 And we are doing detection on each of the frame one by one. 380 00:29:03,690 --> 00:29:07,980 And we are getting the bounding box coordinates unique ID for each of the detected object. 381 00:29:07,980 --> 00:29:10,410 Thus we are getting the class IDs as well. 382 00:29:14,190 --> 00:29:14,580 Okay. 383 00:29:14,580 --> 00:29:15,480 So. 384 00:29:21,980 --> 00:29:26,810 If the length of the output is greater than zero. 385 00:29:42,430 --> 00:29:45,010 The first four are presenting the bounding box coordinates. 386 00:29:46,600 --> 00:29:52,720 Then the second last basically represents the unique ID that we assign to each of the detected object. 387 00:29:57,770 --> 00:30:01,640 And the last one represents the object ID or the class ID. 388 00:30:08,540 --> 00:30:12,530 Okay, so if I just go to the object counter.py file. 389 00:30:12,530 --> 00:30:19,490 So over here you can see that I have created draw boxes function that will draw bounding box coordinates 390 00:30:19,610 --> 00:30:21,200 around each of the detected object. 391 00:30:21,200 --> 00:30:25,070 Plus will add the unique ID that is being assigned to each of the detected object. 392 00:30:25,070 --> 00:30:27,050 Plus will also add the class names. 393 00:30:27,710 --> 00:30:33,350 So I will just write draw dash boxes or I will just write. 394 00:30:34,580 --> 00:30:39,650 So if I just go above over here so we have object counter. 395 00:30:43,470 --> 00:30:45,660 I will assign draw dash boxes. 396 00:30:47,010 --> 00:30:49,380 We have the front frame over here. 397 00:30:50,530 --> 00:30:54,580 Lasts will pass down the warning box or nets. 398 00:30:54,970 --> 00:30:56,440 It was the IDs. 399 00:30:59,580 --> 00:31:00,420 You need guidance. 400 00:31:00,420 --> 00:31:02,970 Plus will class the class ylist here. 401 00:31:03,240 --> 00:31:07,500 Okay, so if I just go to the object counter.py file. 402 00:31:07,500 --> 00:31:12,360 So let me explain you this complete code from the start very start. 403 00:31:12,360 --> 00:31:16,170 So over here you can see that here we have created a function by the name. 404 00:31:16,200 --> 00:31:16,710 Uh okay. 405 00:31:16,710 --> 00:31:18,690 Let's start from the initialization step. 406 00:31:18,690 --> 00:31:22,230 So here I am just creating a line by the name self dot line. 407 00:31:22,380 --> 00:31:22,890 Okay. 408 00:31:23,100 --> 00:31:28,170 So in this uh, dot nine or I am just passing the coordinates for the line. 409 00:31:28,170 --> 00:31:31,110 So let me show you how does it works. 410 00:31:33,590 --> 00:31:34,070 At this. 411 00:31:35,890 --> 00:31:40,810 So for example, this is the video in which I will be doing vehicle scouting. 412 00:31:40,810 --> 00:31:43,240 I will be counting the vehicles entering and leaving. 413 00:31:43,240 --> 00:31:45,700 So I am just creating a line over here. 414 00:31:45,700 --> 00:31:46,480 You can see that. 415 00:31:46,480 --> 00:31:49,300 So when the vehicles passes by this line. 416 00:31:49,300 --> 00:31:54,340 So I will say like this vehicle passes by this time I will say that this vehicle is. 417 00:31:56,520 --> 00:32:00,180 So when this vehicle passes first line, I will say that this vehicle is leaving. 418 00:32:00,570 --> 00:32:01,230 Okay. 419 00:32:01,230 --> 00:32:07,620 And when the vehicle coming from here passes by this line, I will say that this vehicle is entering. 420 00:32:07,650 --> 00:32:09,900 So the vehicle passes by. 421 00:32:09,900 --> 00:32:11,910 I say that vehicle is leaving the lane. 422 00:32:11,910 --> 00:32:18,000 And when the vehicle over here on this lane passes by this line, I will say that vehicle is entering 423 00:32:18,000 --> 00:32:18,450 the lane. 424 00:32:18,450 --> 00:32:18,960 Okay. 425 00:32:19,440 --> 00:32:22,170 So we have this creating a line. 426 00:32:22,170 --> 00:32:23,370 And here I have this. 427 00:32:23,490 --> 00:32:24,900 If you just record this. 428 00:32:24,900 --> 00:32:28,260 So if you just focus over here in that spot where my cursor is. 429 00:32:28,710 --> 00:32:29,340 Okay. 430 00:32:29,340 --> 00:32:36,060 At this point where my question is so you can see if I just go over here, this will be the one to explain 431 00:32:36,060 --> 00:32:39,870 the the first endpoint coordinates of the line. 432 00:32:39,870 --> 00:32:45,870 And so if I just go over here this gave me the second and third point coordinates for this line okay. 433 00:32:45,870 --> 00:32:48,090 So these are the first endpoint coordinates. 434 00:32:48,090 --> 00:32:50,790 And these are the second and third borders for this line. 435 00:32:50,790 --> 00:32:54,120 And you can see those values over here okay. 436 00:32:56,120 --> 00:33:00,740 So you can see that I have this pass on the first endpoint coordinates of the lines values over here. 437 00:33:00,740 --> 00:33:04,040 And these represent the second endpoint coordinates of the line okay. 438 00:33:04,040 --> 00:33:08,300 Plus we are just creating a dictionary over here uh, by the name entering and leaving. 439 00:33:08,300 --> 00:33:10,280 And in the end conditioning dictionary. 440 00:33:10,280 --> 00:33:17,090 We are storing the vehicles or the persons entering, uh, the lane over and thus leaving their dictionary. 441 00:33:17,090 --> 00:33:21,230 We are this will restore the person or vehicles leaving the lane. 442 00:33:21,230 --> 00:33:25,280 And I'm just creating another dictionary by the name Delta Dash DK. 443 00:33:25,280 --> 00:33:31,040 Okay, so here I have just mentioned all the classes name that we have in the Coco data set. 444 00:33:31,040 --> 00:33:36,410 So the Coco or YOLO model has been pre-trained trained on the Coco data set. 445 00:33:36,410 --> 00:33:39,260 And Coco dataset consists of 80 different classes. 446 00:33:39,260 --> 00:33:44,480 So uh, if we have the class id zero, this means it belongs to the person class. 447 00:33:44,480 --> 00:33:47,960 If we have the class id one, it means that it belongs to the bicycle class. 448 00:33:47,960 --> 00:33:50,570 If we have the class 82, it belongs to the car class. 449 00:33:50,570 --> 00:33:56,510 And if we have the class ID four, uh, if we have the class ID three, it belongs to the motorbike 450 00:33:56,510 --> 00:33:56,900 class. 451 00:33:56,900 --> 00:33:57,380 Okay. 452 00:33:58,700 --> 00:34:02,870 And if we have the class ID as 79, then it locks the toothbrush class. 453 00:34:02,870 --> 00:34:07,460 So we start with zero and end at 79. 454 00:34:07,460 --> 00:34:08,060 Okay. 455 00:34:08,060 --> 00:34:12,290 So the class would class ID zero means that it belongs to the person class. 456 00:34:12,290 --> 00:34:14,300 Class ID one means it blocks the bicycle. 457 00:34:14,300 --> 00:34:17,060 Class 82 means blocks of the car class. 458 00:34:17,060 --> 00:34:20,600 Then we have the complete color code labels function. 459 00:34:20,600 --> 00:34:27,350 So if the class ID or label is zero, then it means it belongs to the person class and the bounding 460 00:34:27,350 --> 00:34:29,090 box will have the green color. 461 00:34:29,090 --> 00:34:34,730 If the class ID is two, then it means it belongs to the car class, and the bounding box will have 462 00:34:34,730 --> 00:34:35,720 the following color. 463 00:34:35,720 --> 00:34:40,850 If the class ID is three blocks, the motorbike class and the bounding box will have the the following 464 00:34:40,850 --> 00:34:41,210 color. 465 00:34:41,210 --> 00:34:48,200 If the class ID is five, then it belongs to the bus class and the bounding box will have the following 466 00:34:48,200 --> 00:34:48,440 color. 467 00:34:48,440 --> 00:34:54,620 And if, uh, the class ID doesn't belong to either of these classes, then we will assign that this 468 00:34:54,620 --> 00:34:56,240 following color to the bounding boxes. 469 00:34:56,240 --> 00:35:01,160 Okay, so here I have displayed the initialized Deepsort function where we are just initializing the 470 00:35:01,160 --> 00:35:02,870 Deepsort tracker over here. 471 00:35:02,870 --> 00:35:07,250 And here we have the speed at deepsort configuration of the object. 472 00:35:07,250 --> 00:35:13,430 And we have, uh, where we have loaded all the settings from the Deepsort dot yml file over here. 473 00:35:14,870 --> 00:35:21,650 Okay, so now, uh, what we will say over here is, um, so now here you can see we have the intersect 474 00:35:21,650 --> 00:35:22,190 function. 475 00:35:22,190 --> 00:35:22,640 Okay. 476 00:35:22,640 --> 00:35:24,530 So what does this do? 477 00:35:25,160 --> 00:35:25,730 Okay. 478 00:35:28,640 --> 00:35:35,750 So we are just saying that over here, when the slide, you consider where the center point of the bounding 479 00:35:35,750 --> 00:35:40,940 box, like you can see here, we have the center point of the bounding box and the center point of the 480 00:35:40,940 --> 00:35:41,720 bounding box. 481 00:35:41,720 --> 00:35:48,620 The exact with this line over here, then we will count or we will say that the, uh, volume vehicle 482 00:35:48,620 --> 00:35:49,250 is leaving. 483 00:35:49,250 --> 00:35:55,340 And when the center point of this car coming from here intersect with this line, then we will say the 484 00:35:55,340 --> 00:35:56,690 car is entering the lid. 485 00:35:57,230 --> 00:35:59,630 And when the center point of the. 486 00:36:04,120 --> 00:36:10,840 If the vehicle coming over a car coming from this lane in this sect with this line so that center point 487 00:36:10,840 --> 00:36:16,780 of the vehicle or the car intersect with this line, then we will say, uh, the vehicle is leaving. 488 00:36:16,780 --> 00:36:17,320 Okay. 489 00:36:17,320 --> 00:36:23,650 So when the center point of the vehicle intersect in this line, then you will say a vehicle is leaving. 490 00:36:23,650 --> 00:36:30,790 So if I just open over here, so, like, okay, uh, I can check if two segments intersect. 491 00:36:30,790 --> 00:36:33,730 So there is I found out the solution for the chord over here. 492 00:36:33,730 --> 00:36:38,350 So basically we need to find out how the two lines intersect. 493 00:36:38,350 --> 00:36:41,440 So here we have the intersect chord over here. 494 00:36:41,440 --> 00:36:44,680 So if the two lines intersect so here we have the chord. 495 00:36:44,680 --> 00:36:47,740 So I will be using this piece of code from here. 496 00:36:49,490 --> 00:36:56,180 Okay, so now you can see that I've just added this code over here to find the intersection between 497 00:36:56,180 --> 00:36:59,240 the center point of the vehicle and the line, which I have drawn. 498 00:36:59,270 --> 00:37:01,550 If so. 499 00:37:02,610 --> 00:37:08,160 I will be finding the intersection for the center point of the London line, which have brought it over 500 00:37:08,160 --> 00:37:08,520 here. 501 00:37:10,420 --> 00:37:13,420 So now I have just created a function by the name get prediction. 502 00:37:13,420 --> 00:37:16,210 So uh, what I will say that, uh. 503 00:37:18,410 --> 00:37:20,030 If I just open over here. 504 00:37:20,030 --> 00:37:25,400 So if the vehicle is going into this direction, like this is a north direction. 505 00:37:27,500 --> 00:37:28,130 Okay. 506 00:37:28,340 --> 00:37:30,470 Or if let me just, uh. 507 00:37:31,070 --> 00:37:35,480 So if the vehicle is going into this direction, this is a north direction. 508 00:37:35,480 --> 00:37:36,050 Okay. 509 00:37:36,170 --> 00:37:41,960 And so when the vehicle goes into the north direction, we will say the vehicles are entering. 510 00:37:41,960 --> 00:37:48,290 And if the vehicle goes into this direction which is the south direction, then we will say the vehicles 511 00:37:48,290 --> 00:37:50,150 are leaving the lane. 512 00:37:50,180 --> 00:37:50,720 Okay. 513 00:37:50,720 --> 00:37:55,970 So if the vehicles go into the north direction over here, then you'll say the vehicles are entering. 514 00:37:55,970 --> 00:38:02,300 And if the vehicles go over here, then we will say south direction. 515 00:38:02,300 --> 00:38:03,920 The vehicles are leaving. 516 00:38:03,920 --> 00:38:08,210 And if the vehicles causing the north direction then we'll say the vehicles are entering. 517 00:38:08,750 --> 00:38:09,230 Okay. 518 00:38:11,150 --> 00:38:12,950 So we will find out the direction. 519 00:38:12,950 --> 00:38:15,710 So how we can find out the direction okay. 520 00:38:15,710 --> 00:38:22,940 So let me just tell you how we can find out that direction like in which direction the vehicles are 521 00:38:22,940 --> 00:38:23,660 traveling. 522 00:38:26,010 --> 00:38:28,290 If you go to the drop boxes function. 523 00:38:28,290 --> 00:38:30,630 So over here you can see that we are. 524 00:38:30,840 --> 00:38:34,680 So here you can see that we have created a dictionary by the name data Dash dk. 525 00:38:34,680 --> 00:38:38,070 And uh here you can see that we are just defining it. 526 00:38:38,100 --> 00:38:40,470 DK so DK is basically a list. 527 00:38:40,470 --> 00:38:47,580 Or you can say DK is preferred over list in the cases where we need to quicker in or pop operations. 528 00:38:47,610 --> 00:38:48,030 Okay. 529 00:38:48,030 --> 00:38:52,530 So where we need a quicker in and pop operation where we prefer DK over list. 530 00:38:52,530 --> 00:38:58,770 So DK is just a list and we are just defining the maximum length of this big list will be 64. 531 00:38:58,950 --> 00:38:59,340 Okay. 532 00:38:59,340 --> 00:39:06,090 Over here and in this data DK, we are just appending our center coordinates of the bounding boxes. 533 00:39:06,120 --> 00:39:06,690 Okay. 534 00:39:06,690 --> 00:39:11,340 And in the data DK we are just appending the center coordinates of the bounding boxes. 535 00:39:11,340 --> 00:39:16,890 So let me just, uh, print this out to show you how we will find out the direction. 536 00:39:20,230 --> 00:39:20,560 It. 537 00:39:20,950 --> 00:39:22,930 Let me just run this from the. 538 00:39:25,890 --> 00:39:26,100 It. 539 00:39:28,340 --> 00:39:28,550 The. 540 00:39:33,160 --> 00:39:35,590 So this will take a few seconds. 541 00:39:44,250 --> 00:39:46,260 Um, let us wait for this. 542 00:39:52,650 --> 00:39:52,830 The. 543 00:39:57,030 --> 00:39:57,840 So. 544 00:40:03,050 --> 00:40:03,380 It. 545 00:40:06,740 --> 00:40:08,570 Okay, so I have the data enough. 546 00:40:08,570 --> 00:40:10,310 So let me just show you how the zero. 547 00:40:12,670 --> 00:40:16,120 Okay, so if I just open this up over here. 548 00:40:16,120 --> 00:40:20,170 So now you can see if I just go about about. 549 00:40:33,540 --> 00:40:34,260 So okay. 550 00:40:34,260 --> 00:40:40,530 For here, you can see that, uh, this is basically the unique ID for each of the object which we have. 551 00:40:40,560 --> 00:40:40,980 Okay. 552 00:40:40,980 --> 00:40:44,580 And here we have the dequeue list which you can find over here okay. 553 00:40:44,580 --> 00:40:47,790 And the maximum length of this list is 64 okay. 554 00:40:47,790 --> 00:40:55,710 So now you can see that uh, here we have in the dequeue list, we have more than two different iterations 555 00:40:55,710 --> 00:40:56,250 over here. 556 00:40:56,250 --> 00:41:02,400 Like you can see that if the log here we have added a condition if the length or the data id. 557 00:41:02,400 --> 00:41:09,090 So here you can see that for the ID one the length is equal to two. 558 00:41:09,120 --> 00:41:12,030 Like we have two different iterations over here. 559 00:41:12,540 --> 00:41:15,270 So now what how we can find out the direction okay. 560 00:41:15,270 --> 00:41:21,600 So over here you can see that if the point one is greater than point two okay. 561 00:41:21,600 --> 00:41:28,080 So if the first index of the point one is greater than the first index of the point two okay. 562 00:41:28,080 --> 00:41:31,710 So like this is the first index of the point one. 563 00:41:31,710 --> 00:41:34,020 And this is the first index of the point two. 564 00:41:34,380 --> 00:41:40,380 So over here I'm just passing the data dequeue ID zero and ID one okay. 565 00:41:40,380 --> 00:41:43,080 So this is the data dequeue ID zero. 566 00:41:43,080 --> 00:41:46,410 And this is the data dequeue ID one okay. 567 00:41:46,410 --> 00:41:47,160 So. 568 00:41:48,840 --> 00:41:49,350 Over here. 569 00:41:49,350 --> 00:41:54,780 And we are just finding using the first index which we have defined over here, like we are just considering 570 00:41:54,780 --> 00:41:55,920 out the first index. 571 00:41:55,920 --> 00:42:02,010 So the first index will be this one and this one okay. 572 00:42:02,460 --> 00:42:03,480 So both of these. 573 00:42:03,480 --> 00:42:10,440 So we are saying that if the like here the 783 is greater than 782. 574 00:42:10,440 --> 00:42:14,760 So this means like here you can see that 783 is greater than 782. 575 00:42:14,760 --> 00:42:20,700 So this means the our object or the vehicle or the person is traveling in the south direction. 576 00:42:20,700 --> 00:42:26,670 So if I just show you so this means the vehicle or person is traveling in the south direction which 577 00:42:26,670 --> 00:42:28,770 means the vehicle is leaving. 578 00:42:30,120 --> 00:42:30,540 Okay. 579 00:42:30,690 --> 00:42:37,440 If this, uh, point one is less than point two, like here, the country, the point one is greater 580 00:42:37,440 --> 00:42:38,040 than point two. 581 00:42:38,040 --> 00:42:40,410 So this is traveling the south south direction. 582 00:42:40,410 --> 00:42:44,340 But, uh, let me consider another case where this is happening. 583 00:42:44,340 --> 00:42:44,970 Okay. 584 00:42:45,630 --> 00:42:47,040 Like here you can see that. 585 00:42:47,040 --> 00:42:54,450 So 137 so here you can see that point one is less than point two. 586 00:42:54,480 --> 00:42:59,970 So that you can say that vehicle or the person is traveling in the north direction okay. 587 00:43:00,150 --> 00:43:00,900 So. 588 00:43:05,100 --> 00:43:08,370 What if we say the vehicle over the person is traveling the north direction? 589 00:43:08,370 --> 00:43:12,540 Then we are saying that the person is entering or the vehicle is entering. 590 00:43:12,540 --> 00:43:16,860 When we say the person or the vehicle is traveling the south direction, we can we are saying that the 591 00:43:16,860 --> 00:43:18,510 person or the vehicle is leaving. 592 00:43:18,540 --> 00:43:20,970 Okay, so how we can find out. 593 00:43:20,970 --> 00:43:26,430 Like you can see over here, 783 is greater than 782. 594 00:43:26,430 --> 00:43:32,970 So we are just saying that it's traveling in the South direction because 783 represents the latest represent 595 00:43:32,970 --> 00:43:34,140 the latest iteration. 596 00:43:34,140 --> 00:43:35,700 And this is the previous side. 597 00:43:36,000 --> 00:43:40,440 Like uh so here this is one represent the latest iteration. 598 00:43:40,440 --> 00:43:40,890 Okay. 599 00:43:41,190 --> 00:43:46,650 So if I just open this over here like you can see over here, this is the starting .00. 600 00:43:46,650 --> 00:43:49,680 And at this extreme point we have the higher value. 601 00:43:49,680 --> 00:43:52,980 So the maximum point of the width and the height is basically this point. 602 00:43:52,980 --> 00:43:54,180 This end point okay. 603 00:43:54,510 --> 00:43:59,820 So when we say that 783 is greater than 782 okay. 604 00:43:59,820 --> 00:44:00,750 So. 605 00:44:01,840 --> 00:44:03,670 Like you can see out here. 606 00:44:04,390 --> 00:44:09,520 Uh, if I just see, say, over here, that's 783 is greater than 782. 607 00:44:09,820 --> 00:44:15,190 So this basically means that we are traveling on the higher side over here. 608 00:44:15,220 --> 00:44:15,940 Okay. 609 00:44:16,030 --> 00:44:18,520 Because the previous was 783. 610 00:44:19,060 --> 00:44:20,680 Uh, the previous was 782. 611 00:44:20,680 --> 00:44:23,200 And now the latest was 783. 612 00:44:23,200 --> 00:44:25,450 So we are traveling in the south direction. 613 00:44:25,450 --> 00:44:25,900 Okay. 614 00:44:25,930 --> 00:44:29,410 The higher side, because here we have the maximum weight and the height. 615 00:44:29,410 --> 00:44:29,770 Okay. 616 00:44:29,830 --> 00:44:31,690 And when I say that. 617 00:44:33,780 --> 00:44:35,280 Now when I say that over here. 618 00:44:37,050 --> 00:44:44,160 So like you can see over here, I can say over here where the previous point. 619 00:44:47,720 --> 00:44:51,080 Oh, let me consider the case where we are traveling in the north direction. 620 00:44:56,000 --> 00:44:56,420 Like me. 621 00:44:56,420 --> 00:44:57,830 They are just so over here. 622 00:44:57,830 --> 00:45:01,460 Like you can see that this is 137 and 136. 623 00:45:01,460 --> 00:45:01,790 Okay. 624 00:45:01,790 --> 00:45:07,730 So the latest iteration is 136 and the previous iteration was uh, 137. 625 00:45:07,730 --> 00:45:08,180 Okay. 626 00:45:08,180 --> 00:45:15,650 So when there is 136 uh, so we are uh, just moving towards the north direction because in the over 627 00:45:15,650 --> 00:45:19,040 towards this, our height will be decreasing. 628 00:45:19,040 --> 00:45:19,490 Okay. 629 00:45:19,490 --> 00:45:23,060 So towards this our height will be decreasing in the north direction. 630 00:45:23,060 --> 00:45:25,160 Our height will be decreasing okay. 631 00:45:25,160 --> 00:45:29,540 And thus our direction, our height will be increasing like in the south direction. 632 00:45:29,540 --> 00:45:32,000 We will be, uh, going towards our maximum height. 633 00:45:32,000 --> 00:45:36,590 And in the north direction our height will be continuously decreasing. 634 00:45:36,590 --> 00:45:37,040 Okay. 635 00:45:39,140 --> 00:45:45,410 So like you can see over here, we have 137 and 136 for the latest iteration is 136. 636 00:45:45,410 --> 00:45:51,350 It means that we are going in the north direction and our height is constantly continuously decreasing. 637 00:45:51,350 --> 00:45:53,630 So here the vehicles are entering. 638 00:45:53,780 --> 00:45:56,960 So I have just added the same word over here. 639 00:45:56,960 --> 00:45:58,310 Is the south in direction. 640 00:45:58,310 --> 00:46:01,460 So the vehicles are leaving or the person is leaving. 641 00:46:01,460 --> 00:46:04,700 And if the north is direction then the person is entering over. 642 00:46:04,700 --> 00:46:05,990 The vehicles are entering. 643 00:46:05,990 --> 00:46:07,520 So the same code is being added. 644 00:46:07,520 --> 00:46:09,350 And here we are just finding the intersection. 645 00:46:09,350 --> 00:46:11,270 So when we say that uh. 646 00:46:13,090 --> 00:46:15,670 The center point of this car and this, like, intersect. 647 00:46:16,120 --> 00:46:16,510 Okay. 648 00:46:16,510 --> 00:46:18,940 So then we will just get better direction. 649 00:46:18,940 --> 00:46:25,840 In which direction the car or the vehicle is being, uh, the car, vehicle or person is being traveling. 650 00:46:25,840 --> 00:46:26,350 Okay. 651 00:46:26,350 --> 00:46:32,260 And here I've just displayed the count of above like the leaving count or entering count over here. 652 00:46:32,260 --> 00:46:39,550 So let me just run this complete script file and show you what output are we getting from here? 653 00:46:43,490 --> 00:46:44,510 Oh, let me just. 654 00:46:52,760 --> 00:46:53,030 It. 655 00:46:53,030 --> 00:46:55,910 So now we will be running on the. 656 00:47:00,070 --> 00:47:00,670 Okay. 657 00:47:00,700 --> 00:47:05,980 And I will just write you that image so that I can show you the output as well. 658 00:47:06,190 --> 00:47:08,800 And here we will be using a text box. 659 00:47:08,920 --> 00:47:11,950 For now we will be using S3 dot mp4. 660 00:47:11,950 --> 00:47:12,280 So. 661 00:47:42,410 --> 00:47:47,210 So now you can see that currently we don't like you can see the line drawn over here. 662 00:47:47,420 --> 00:47:49,070 And this is the wrong detection. 663 00:47:49,100 --> 00:47:49,700 Okay. 664 00:47:49,700 --> 00:47:53,750 So now you can see that here we have the center point of this car. 665 00:47:53,780 --> 00:47:55,220 You can see this green line. 666 00:47:55,640 --> 00:47:56,270 Okay. 667 00:47:56,690 --> 00:47:59,960 And here you can see the center point of this car as well. 668 00:48:03,640 --> 00:48:03,940 It. 669 00:48:08,730 --> 00:48:10,710 So let's go ahead. 670 00:48:10,710 --> 00:48:14,940 When the center point intersect with this line, then we will say the vehicle is entering. 671 00:48:14,940 --> 00:48:19,470 And when this center point intersect with this line, then we can say that the car is leaving. 672 00:48:19,590 --> 00:48:19,980 Okay. 673 00:48:29,220 --> 00:48:32,130 So let me just show you how this works. 674 00:48:34,720 --> 00:48:40,330 Okay so I'm using CPU so you can see the processing or the frame rate is very low. 675 00:48:57,050 --> 00:49:00,230 But when the center point intersects with this line, then we will see. 676 00:49:18,490 --> 00:49:23,350 So now you can see the center point as the center point intersect with this line. 677 00:49:24,010 --> 00:49:24,640 Okay. 678 00:49:24,940 --> 00:49:26,620 So the color of the line drawn wide. 679 00:49:26,620 --> 00:49:30,070 And you can see that we have the entering count of the car as one. 680 00:49:30,610 --> 00:49:31,210 Okay. 681 00:49:32,740 --> 00:49:38,020 And when this center point intersect with this line then we will have the leaving count of the car as 682 00:49:38,020 --> 00:49:38,710 one. 683 00:49:46,010 --> 00:49:50,660 And you can say that using a we are doing integrating object detection with object tracking and using 684 00:49:50,660 --> 00:49:51,800 Deepsort algorithm. 685 00:49:51,800 --> 00:49:57,350 Like you can see that we have assigned a unique ID seven to this object to this car, we have assigned 686 00:49:57,350 --> 00:49:58,850 a unique ID two to this car. 687 00:49:58,850 --> 00:50:01,250 We have assigned a unique ID 15 to this car. 688 00:50:01,370 --> 00:50:03,680 We have assigned a unique ID 16 to this car. 689 00:50:03,680 --> 00:50:06,770 We have assigned a unique ID to this type card. 690 00:50:06,770 --> 00:50:08,750 So this is what we do in object tracking. 691 00:50:08,750 --> 00:50:13,100 Now the object tracking we assign a unique value to each of the detected object. 692 00:50:13,100 --> 00:50:16,970 And then we track that object throughout the entire frame okay. 693 00:50:39,420 --> 00:50:45,870 Oh, I'm just waiting for this car to intersect with this line so that I can show you the living count 694 00:50:45,870 --> 00:50:47,190 over here as well. 695 00:50:52,840 --> 00:50:56,830 Now you can see that as the center point intersect with this line. 696 00:50:56,830 --> 00:51:00,250 Then we will see uneven count displayed over here as well. 697 00:51:12,860 --> 00:51:15,020 Okay, so now you can see the living count of card. 698 00:51:15,050 --> 00:51:20,660 So in this way you can count the vehicles entering and leaving our lane as well. 699 00:51:20,660 --> 00:51:23,930 So I will just test on other video as well. 700 00:51:23,930 --> 00:51:33,470 So before testing on other video like for the person counting, uh, let me just make one change the. 701 00:51:41,590 --> 00:51:47,650 So I'm just so here that different coordinates with persuading the video frame. 702 00:51:47,950 --> 00:51:48,310 Okay. 703 00:51:48,310 --> 00:51:56,290 So plus over here the line coordinates will also change or the dimensions of the line will also change. 704 00:51:56,290 --> 00:52:01,150 These are the starting point and this is the ending point of the line if. 705 00:52:03,540 --> 00:52:07,530 And now I will just writing test walk. 706 00:52:07,560 --> 00:52:08,250 Okay. 707 00:52:08,850 --> 00:52:10,950 And this is what the person. 708 00:52:11,220 --> 00:52:20,850 And so let's run this and let's see if you are able to do person counting and entering or leaving lane 709 00:52:20,850 --> 00:52:21,600 or not. 710 00:52:24,350 --> 00:52:24,650 If. 711 00:52:37,210 --> 00:52:37,810 Oh, yeah. 712 00:52:37,810 --> 00:52:41,050 You can see we have the video over here. 713 00:52:44,980 --> 00:52:45,250 Yet. 714 00:52:51,970 --> 00:52:52,270 Rod. 715 00:52:52,270 --> 00:52:54,940 You can see here we have the lane, the green lane. 716 00:52:54,940 --> 00:53:02,110 So when this person intersect with this line we will see a leaving cone displayed over here. 717 00:53:18,710 --> 00:53:19,820 So let's see. 718 00:53:19,820 --> 00:53:20,240 Order. 719 00:53:22,310 --> 00:53:27,890 So as this person intersect with this line and we'll see a living cone. 720 00:53:27,890 --> 00:53:30,710 And when this person intersect with this line we will see entering. 721 00:53:30,830 --> 00:53:33,680 So this person is traveling in the north direction. 722 00:53:33,680 --> 00:53:36,260 And this person is traveling a south direction. 723 00:53:36,260 --> 00:53:38,720 So in the south direction we'll have the living cow. 724 00:53:38,720 --> 00:53:41,060 And in the north direction we will have the entering cow. 725 00:53:47,240 --> 00:53:47,480 A. 726 00:53:52,070 --> 00:54:00,020 And you can see using object tracking, we have assigned a unique ID 517, 21 2519 1013 eight to each 727 00:54:00,020 --> 00:54:00,440 person. 728 00:54:00,440 --> 00:54:01,400 So let's see. 729 00:54:02,510 --> 00:54:03,560 Okay. 730 00:54:07,130 --> 00:54:10,070 Okay, so now you can see we have the entering out over here. 731 00:54:10,070 --> 00:54:14,900 And as this person intersect with this line we will have the leaving out should be over here. 732 00:54:15,860 --> 00:54:19,220 So now you can see the center point intersect with this line. 733 00:54:19,760 --> 00:54:23,990 Now you can see that we have the leaving count because this person have also crossed this line. 734 00:54:23,990 --> 00:54:25,940 So two person simultaneously crossed. 735 00:54:25,970 --> 00:54:30,350 We have the leaving count of person two and the entering count of the person as one. 736 00:54:30,860 --> 00:54:36,920 So now you can see over here we are able to complete the application where we are doing persons counting, 737 00:54:36,920 --> 00:54:41,480 entering and leaving a lane over the vehicles, counting and finger leaving a lane. 738 00:54:41,480 --> 00:54:42,920 So that's all from this tutorial. 739 00:54:42,920 --> 00:54:43,910 Thank you for watching.