1 00:00:01,340 --> 00:00:01,970 Hi guys. 2 00:00:02,270 --> 00:00:07,520 So we are ready to develop the installing algorithm and the very first date that we'll be developing 3 00:00:07,670 --> 00:00:11,000 will be the localization states now inside the state. 4 00:00:11,270 --> 00:00:14,870 The primary goal is to localize our robot inside the robot. 5 00:00:15,200 --> 00:00:21,110 For this we have the Salvador part where we have the missiles asylum function that will be used iteratively 6 00:00:21,200 --> 00:00:24,290 on all of those stages once they have been developed. 7 00:00:24,660 --> 00:00:28,670 Now, because we are working on the localization states and we haven't developed it yet. 8 00:00:28,910 --> 00:00:32,690 So first we need to develop it before calling the asylum asylum function. 9 00:00:33,230 --> 00:00:39,470 So let's go inside the maze port and create a new that we named as bot localization. 10 00:00:41,640 --> 00:00:44,520 But firefighters are inside it. 11 00:00:44,790 --> 00:00:49,890 The very first thing that we need to do is to simply import the two most important libraries that are 12 00:00:49,890 --> 00:00:51,600 civil to by library. 13 00:00:51,870 --> 00:00:58,010 Now CV two is to import the computer vision algorithms, and numpy is great for creating an empire. 14 00:00:58,770 --> 00:01:04,950 Then we create the class of mortal collider that will perform the actual task of localizing a robot. 15 00:01:05,670 --> 00:01:10,050 The very first function that we'll be writing inside this class will be the initialization function. 16 00:01:10,590 --> 00:01:16,200 So that can be done by simply writing definition in its first argument of self. 17 00:01:16,680 --> 00:01:18,810 And then we define is instance variable. 18 00:01:19,110 --> 00:01:24,460 The first instance variable is a boolean variable that simply tracks whether we have extracted the background 19 00:01:24,540 --> 00:01:24,810 or not. 20 00:01:25,240 --> 00:01:28,200 Because this is the Initialize and states, we haven't extracted it. 21 00:01:28,530 --> 00:01:34,290 So we set this to false and then we create a placeholder for keeping our background model. 22 00:01:34,500 --> 00:01:38,520 Once we have expected this, so we initialize this to an empty list. 23 00:01:39,060 --> 00:01:41,420 Once you've done this, you move on to the next function. 24 00:01:41,430 --> 00:01:43,650 That will be our main function for this. 25 00:01:43,650 --> 00:01:47,430 Localization states that will be known as localized board function. 26 00:01:47,820 --> 00:01:52,410 This takes in the input argument of self and the second argument of grand frame. 27 00:01:52,680 --> 00:01:54,510 The third argument of frame to display. 28 00:01:54,990 --> 00:02:00,000 Now frame display is be simply used to display the board that has been localized on the frame. 29 00:02:00,690 --> 00:02:06,120 Now we defined this part of a function that is that we check whether we have extracted or whether we 30 00:02:06,120 --> 00:02:08,840 have performed the full step of our location state. 31 00:02:09,060 --> 00:02:11,520 That is whether we have extracted the background or not. 32 00:02:11,940 --> 00:02:14,160 So that can be done by simply writing, if not. 33 00:02:15,380 --> 00:02:18,020 So now it is background obstructed or not. 34 00:02:18,350 --> 00:02:23,270 Now, if it has not been extracted, then we simply call the function that extracts the background. 35 00:02:24,590 --> 00:02:29,630 So there is no one to extract background in providing the current frame as an input. 36 00:02:30,050 --> 00:02:36,950 And then we simply after that we set the boolean to true because the background has been extracted successfully. 37 00:02:37,700 --> 00:02:42,170 Now this means the frame to implement this particular function of extract background. 38 00:02:42,710 --> 00:02:43,430 Now let's do that. 39 00:02:43,970 --> 00:02:45,950 So we go up right here. 40 00:02:45,960 --> 00:02:46,730 The definition. 41 00:02:47,210 --> 00:02:48,260 Extract background. 42 00:02:48,770 --> 00:02:52,100 Providing the first argument of self and the second argument of frame. 43 00:02:53,060 --> 00:02:58,760 Now inside it, the very first thing that you need to do is to simply extract the mass of all the region 44 00:02:58,760 --> 00:03:00,650 of interest that are present inside the frame. 45 00:03:01,250 --> 00:03:06,860 So extract the most of all the region of interest. 46 00:03:07,250 --> 00:03:08,030 Now, what are those? 47 00:03:08,030 --> 00:03:08,780 Region of interest? 48 00:03:09,080 --> 00:03:10,880 There's the Miz and there's the car. 49 00:03:11,870 --> 00:03:17,060 So how do we extract from the video we are going to extract them is using as detection algorithm. 50 00:03:17,390 --> 00:03:22,970 Now edges are only present for those objects that have sharp features around them. 51 00:03:23,450 --> 00:03:24,650 So let's do that. 52 00:03:24,890 --> 00:03:30,800 So the way we do that is using the detecting algorithm and the cutting edge detection algorithm first 53 00:03:30,800 --> 00:03:32,090 requires a grayscale image. 54 00:03:32,210 --> 00:03:36,920 So we need to convert our RGV or BTR color frame to a grayscale image. 55 00:03:37,310 --> 00:03:42,650 And that can be done by simply writing simple to convert color, providing the frame as the input. 56 00:03:43,870 --> 00:03:45,820 And Stevie to color. 57 00:03:47,570 --> 00:03:48,250 BGR. 58 00:03:50,090 --> 00:03:52,550 To Ray as the court. 59 00:03:53,300 --> 00:04:01,430 This gives us the output of break and then for the next we simply need to pass this grayscale to the 60 00:04:01,430 --> 00:04:02,720 cancer detection algorithm. 61 00:04:03,890 --> 00:04:10,100 Now can you takes an image of grayscale, then we provided the range of 50 to 150. 62 00:04:10,940 --> 00:04:13,850 Now this range provides adequate edges if we want. 63 00:04:14,810 --> 00:04:17,630 Then for the next argument for edges, we provide none. 64 00:04:17,990 --> 00:04:20,180 And for upward your size, we provide three. 65 00:04:21,080 --> 00:04:26,660 Now, after applying the scanning detection algorithm to get edges as output, once we have the edges, 66 00:04:26,720 --> 00:04:31,970 we can now use the open CV function of fine contours to find the contours of all those regions of interest. 67 00:04:32,510 --> 00:04:40,520 So we use the function of fine controls that takes input, argument, image that is ideal for the mode 68 00:04:40,520 --> 00:04:48,230 we provide in CV to return external, to simply get the external contours of all those objects and for 69 00:04:48,230 --> 00:04:48,770 the method. 70 00:04:50,720 --> 00:04:58,910 We provide in Syria to saint approximation none now because this function of plus two or two gives to 71 00:04:58,910 --> 00:05:02,330 outputs and the only one, the first output that is the control. 72 00:05:02,540 --> 00:05:04,250 So we need to write here zero. 73 00:05:05,170 --> 00:05:08,800 So this will give us the output of controls. 74 00:05:10,540 --> 00:05:15,880 And once we have the controls, we can simply now loop over these controls and simply draw them on an 75 00:05:15,880 --> 00:05:16,990 empty placeholder. 76 00:05:17,410 --> 00:05:20,290 Now, at the moment, we have not defined that placeholder. 77 00:05:20,320 --> 00:05:24,660 So let's define a placeholder because we were expecting the region of interest for the most intimate 78 00:05:24,760 --> 00:05:25,450 region of interest. 79 00:05:25,450 --> 00:05:28,990 Must we use the function of number zero? 80 00:05:29,800 --> 00:05:33,580 We need to provide a ship that can be taken from the shape of the frame. 81 00:05:35,900 --> 00:05:44,180 So you provide the shape of the frame, except that you won't want the dare type of the frame because 82 00:05:44,180 --> 00:05:46,130 the airframe was of the dirty type. 83 00:05:47,570 --> 00:05:49,360 BGR and we want to get it out. 84 00:05:49,550 --> 00:05:50,480 There it is, Chris. 85 00:05:50,600 --> 00:05:52,610 That is off the mark. 86 00:05:53,180 --> 00:05:54,380 So we tried d-type. 87 00:05:56,500 --> 00:05:59,710 Of number know and it. 88 00:06:02,720 --> 00:06:06,350 So once you've done this, we have the region of interest miles from the place order created. 89 00:06:06,710 --> 00:06:10,520 Now we can simply go all in or simply call it a loop. 90 00:06:11,000 --> 00:06:15,260 It loops over all those contours and gives us the ideas of these controls. 91 00:06:15,830 --> 00:06:16,970 And that can be done for. 92 00:06:16,970 --> 00:06:24,350 And by calling enumerate function, depositing the controls as the iterable, and then simply calling 93 00:06:24,350 --> 00:06:32,060 the function of CB to draw controls and draw controls takes in the input argument of the image, which 94 00:06:32,060 --> 00:06:34,340 is our placeholder that we have created just now. 95 00:06:34,790 --> 00:06:36,080 That is a region of interest, Miles. 96 00:06:36,960 --> 00:06:46,590 For the controls group Paulson controls and for an index forcing iterable that is idea and for the color 97 00:06:46,590 --> 00:06:53,250 three passing to 55 and for thickness we write negative one receives to fill all those controls. 98 00:06:53,760 --> 00:06:59,820 Now what this function will do was simply draw all of the mask of all those regional interests that 99 00:06:59,820 --> 00:07:00,950 we are just detected now. 100 00:07:01,710 --> 00:07:05,970 Once you've done this, you've completed the step of expecting the mask of all the region of interest 101 00:07:06,360 --> 00:07:07,110 for the next step. 102 00:07:07,350 --> 00:07:15,090 What we need to do is to simply remove the guard from this region of interest, because although it 103 00:07:15,090 --> 00:07:19,920 is a region of interest, but this is not part of the background, the guard is the part of a foreground. 104 00:07:20,190 --> 00:07:24,780 So to get the background or to extract the beginning, we need to remove the car from the background. 105 00:07:25,260 --> 00:07:30,180 And the way we can do this is simply leveraging the information that God is the smallest object on the 106 00:07:30,180 --> 00:07:30,480 scene. 107 00:07:30,960 --> 00:07:33,520 So we need to find the object this one has gone to. 108 00:07:34,110 --> 00:07:40,140 So for this, we need to use a function that we have created inside the utilities folder, the utilities 109 00:07:40,140 --> 00:07:41,820 file that we have created earlier. 110 00:07:42,510 --> 00:07:46,710 So let's import that file and use the function for the smallest object. 111 00:07:47,340 --> 00:07:49,560 So we need to import it right here. 112 00:07:50,640 --> 00:07:57,950 Import not you to or from from go to utilities to import the returns. 113 00:07:57,960 --> 00:07:58,860 Smallest object. 114 00:08:01,460 --> 00:08:06,440 So we go right here and simply call the function that we have just important files in the controls, 115 00:08:06,590 --> 00:08:09,500 which gives us the output of index of the smallest title. 116 00:08:09,920 --> 00:08:13,400 So we named that as minimum control index. 117 00:08:13,730 --> 00:08:19,490 And once we have this, we can now use this index and to eliminate the car from all those region of 118 00:08:19,490 --> 00:08:21,080 interest that we have extracted earlier. 119 00:08:21,620 --> 00:08:27,410 So we don't want to overwrite over the region of interest lost, but we need to create a new placeholder. 120 00:08:27,590 --> 00:08:31,850 So we name that placeholder as a region of interest with no card. 121 00:08:33,200 --> 00:08:38,600 So this is a region of interest, no card models, and this will be a copy of the Region of Interest 122 00:08:38,600 --> 00:08:39,050 Master. 123 00:08:40,520 --> 00:08:41,630 Just in a different place. 124 00:08:42,050 --> 00:08:46,400 Once we have this, we check whether the minimum control index was legit or not. 125 00:08:46,910 --> 00:08:51,320 So if it was legit, it will not be equal to negative one. 126 00:08:51,560 --> 00:08:53,240 Meaning we have found this one as control. 127 00:08:54,800 --> 00:08:55,910 Which is actually the car. 128 00:08:57,830 --> 00:08:58,160 So. 129 00:08:59,270 --> 00:09:03,890 If it is not equal to negative one, then what we do to you to see to draw control function. 130 00:09:05,940 --> 00:09:08,370 To draw over the image that is a region of interest. 131 00:09:08,460 --> 00:09:12,740 No car must provide in the controls as the controls for the index. 132 00:09:12,740 --> 00:09:20,220 They will not provide the minimum contour index because we have to do it on the models of our car and 133 00:09:20,220 --> 00:09:25,890 because we are removing the car from all the region of interest, we need to write here zero as the 134 00:09:25,890 --> 00:09:26,230 color. 135 00:09:26,430 --> 00:09:29,700 So zero will simply remove the car as a region of interest. 136 00:09:30,240 --> 00:09:34,500 And because we want to complete the complete car, not just as we need to. 137 00:09:34,510 --> 00:09:34,740 Right. 138 00:09:34,740 --> 00:09:35,490 Has negative one. 139 00:09:36,030 --> 00:09:38,970 Once you've done this, we have removed the car from our region of interest. 140 00:09:39,540 --> 00:09:45,060 Now, we also require the models of the car separately from me to create a new placeholder. 141 00:09:45,090 --> 00:09:50,010 So we create a new class folder that we name as car models and then draw that over the car models. 142 00:09:50,280 --> 00:09:56,520 The mean control index by providing the color of 25 and then diluting it by drawing with a thickness 143 00:09:56,520 --> 00:09:56,940 of three. 144 00:09:57,390 --> 00:09:59,640 Once we have done that, we can move on to the next step. 145 00:10:00,030 --> 00:10:06,150 Now, because we have the car models will be required area that does not include the car or does not 146 00:10:06,150 --> 00:10:07,260 include the car region. 147 00:10:07,680 --> 00:10:14,190 So that can be done by simply using the function bitwise not not bitwise not simply inverse what ever 148 00:10:14,200 --> 00:10:16,560 was one zero and 0 to 1. 149 00:10:16,980 --> 00:10:21,060 So we provide in the car models to get the opposite of the car models for the region. 150 00:10:21,060 --> 00:10:22,260 That does not include the car. 151 00:10:23,880 --> 00:10:28,860 So that can be done by simply writing a car list or not. 152 00:10:28,860 --> 00:10:29,610 Car must. 153 00:10:33,460 --> 00:10:36,670 This means that we have the region that does not include the car. 154 00:10:36,970 --> 00:10:38,320 So we can use this mosque. 155 00:10:38,500 --> 00:10:41,580 It does not include the car to extract the frame. 156 00:10:41,650 --> 00:10:42,940 That does not include the car. 157 00:10:43,420 --> 00:10:49,720 So we can now extract the frame where the car is removed by simply using the bitwise in. 158 00:10:54,950 --> 00:10:58,970 Which wasn't providing the frame a source one and source two. 159 00:10:59,540 --> 00:11:03,650 And then for the mask and now providing the not guard mask. 160 00:11:05,440 --> 00:11:08,450 Now this simply says only get those region of the frame. 161 00:11:08,470 --> 00:11:09,910 That does not include the car. 162 00:11:10,510 --> 00:11:16,450 And this becomes the frame the car remote now because we have the frame with the car removed, we have 163 00:11:16,450 --> 00:11:18,460 an empty region inside the frame. 164 00:11:18,640 --> 00:11:21,130 So we cannot say that this is the background as of yet. 165 00:11:21,400 --> 00:11:24,670 We have to fill that empty region with the ground replica. 166 00:11:25,210 --> 00:11:31,010 So for the replica we simply take the color of the frame or the top left corner of the frame removed. 167 00:11:31,720 --> 00:11:33,070 So let's do that. 168 00:11:35,110 --> 00:11:38,290 So again, the frame removed and it's top left corner. 169 00:11:40,050 --> 00:11:40,380 First. 170 00:11:42,740 --> 00:11:46,340 And named that as the best color on this fiscal. 171 00:11:46,340 --> 00:11:51,740 It would be used to create the brown replica that would fill in the empty space inside the area where 172 00:11:51,740 --> 00:11:52,790 the car was removed from. 173 00:11:53,540 --> 00:11:55,070 So this cleared the ground replica 174 00:11:58,280 --> 00:12:02,390 and that can be created by simply using the number function of ones like. 175 00:12:04,780 --> 00:12:11,620 This takes an input argument of frame because we want a colour image the same size of the frame. 176 00:12:12,160 --> 00:12:16,510 But we want to multiply this with a base color to get the ground replica. 177 00:12:16,840 --> 00:12:23,020 Now that we have the ground replica, we can use this information to get only the ground region where 178 00:12:23,020 --> 00:12:25,630 the car was present and then fill it with the frame car. 179 00:12:26,200 --> 00:12:32,470 And that can be done by simply using a few of the bitwise and where we get the ground replica only in 180 00:12:32,470 --> 00:12:37,600 the region where the car was present and then simply or in that region with the frame removed. 181 00:12:37,780 --> 00:12:39,400 This gives us the background model. 182 00:12:39,610 --> 00:12:41,560 And now we have extracted the background model. 183 00:12:41,920 --> 00:12:44,230 The main task of this extracted background function. 184 00:12:44,500 --> 00:12:49,420 The main goal of this particular localized state was to localize the robot inside the mix. 185 00:12:49,900 --> 00:12:50,710 So we need. 186 00:12:50,980 --> 00:12:53,830 So we want to localize our robot relative to the Miz. 187 00:12:54,190 --> 00:12:57,820 So to get to the relative home is really quite a frame of reference. 188 00:12:58,120 --> 00:13:02,800 So we need to get that frame of reference or get that means cropped out from the whole frame. 189 00:13:03,310 --> 00:13:07,810 So we require some sort of boundary enclosing that means. 190 00:13:08,320 --> 00:13:09,520 So let's do that. 191 00:13:10,500 --> 00:13:10,690 Okay. 192 00:13:11,390 --> 00:13:17,300 So for that, we first need to create a function that encloses all the region of the means and the car 193 00:13:17,300 --> 00:13:18,050 inside of it. 194 00:13:18,590 --> 00:13:21,260 So we create a function that does exactly that. 195 00:13:21,950 --> 00:13:25,040 So that function would be known as return region of interest bounding. 196 00:13:25,520 --> 00:13:31,130 So we provide in the region of interest models and control, and this will give us the whole that encloses 197 00:13:31,130 --> 00:13:32,870 all of the mess inside of it. 198 00:13:34,270 --> 00:13:41,470 So we simply now call that particular function by simply writing self interest, bounding or providing 199 00:13:41,470 --> 00:13:43,330 the region of interest and the controls. 200 00:13:43,690 --> 00:13:48,130 This gives us the whole that encloses all of those region of interest and includes the means in the 201 00:13:48,130 --> 00:13:48,460 car. 202 00:13:49,150 --> 00:13:54,760 And once we have the whole, we simply use the bounding rect to get the rectangle that bounds, that 203 00:13:54,760 --> 00:13:55,090 hope. 204 00:13:55,480 --> 00:14:00,550 This gives us the starting rule, starting column width and the height of the bounding rectangle. 205 00:14:01,090 --> 00:14:05,470 Once we have that, we use that region to crop out the mes from the whole frame. 206 00:14:05,950 --> 00:14:12,280 So we get the region of interest with no corners and use a crop of the maize from that particular region 207 00:14:12,430 --> 00:14:13,510 or that particular mass. 208 00:14:13,930 --> 00:14:16,510 Once we have the mis cropped out, we can now. 209 00:14:17,540 --> 00:14:22,400 Get them is occupancy grade where the car contributes by simply inverting the is. 210 00:14:23,590 --> 00:14:28,480 And the reason you're inverting them is, is because that region with the males was present or the males 211 00:14:28,480 --> 00:14:32,740 boys were present is directly opposite to the region where the card controls from. 212 00:14:33,070 --> 00:14:37,840 So right now we have the was detected, but we want the opposite area to the wall. 213 00:14:38,020 --> 00:14:47,050 So we need to use CV to bitwise not so we use that which was not to get the occupancy group. 214 00:14:48,740 --> 00:14:50,720 This provides this to me documents secret. 215 00:14:51,140 --> 00:14:57,920 Now this gives us the maze occupancy, but we also want it to be a specific frame of reference. 216 00:14:58,280 --> 00:15:04,430 So to maintain the consistency in the frame of reference, we always point entry point of a maze on 217 00:15:04,430 --> 00:15:04,790 the top. 218 00:15:05,120 --> 00:15:08,630 So because the entry point of the maze of this maze is on the right. 219 00:15:08,630 --> 00:15:10,760 So we need to rotate it counterclockwise. 220 00:15:11,180 --> 00:15:14,450 So let's take it to get the final maze occupancy grid. 221 00:15:14,990 --> 00:15:19,940 And that can be done by switch to rotate, providing the maze occupancy rate. 222 00:15:25,810 --> 00:15:30,100 And for the code we provided Siri to rotate. 223 00:15:31,790 --> 00:15:33,680 Counterclockwise, 90 degrees. 224 00:15:34,040 --> 00:15:38,810 And this gives us so of May's occupancy. 225 00:15:38,830 --> 00:15:39,080 Great. 226 00:15:41,100 --> 00:15:41,280 Now. 227 00:15:41,280 --> 00:15:42,540 We also need to create this. 228 00:15:44,430 --> 00:15:47,820 Inside the function for declared distance in function. 229 00:15:49,190 --> 00:15:53,090 By writing these documents include initializing it with an empty list. 230 00:15:53,540 --> 00:15:58,370 Now this will the second output that we required, and that also become crucial in the next stage for 231 00:15:58,370 --> 00:15:59,120 the mapping stage. 232 00:15:59,900 --> 00:16:05,210 Now, once we have done this, all that is left for us to do it will simply update the parameters that 233 00:16:05,210 --> 00:16:07,400 we used to get this frame of reference. 234 00:16:07,670 --> 00:16:09,410 That is our maze occupancy grid. 235 00:16:09,890 --> 00:16:13,010 That includes the cropping barometers and the rotation parameters. 236 00:16:13,520 --> 00:16:15,920 So let's update those parameters. 237 00:16:16,190 --> 00:16:19,700 So first we need to do is create a few more instance variables. 238 00:16:20,150 --> 00:16:26,180 This includes the growth parameters X, Y and row and column of the original frame and the rotation 239 00:16:26,840 --> 00:16:33,320 angle that that was used to rotate out of our frame to get the maze entry on the top. 240 00:16:34,130 --> 00:16:40,850 Now we need to create a new function that would be used to update the frame of reference parameters. 241 00:16:41,240 --> 00:16:45,350 So we use this particular function by calling it right below here. 242 00:16:45,380 --> 00:16:50,900 So we call that particular function of or update frame of reference provided stock roll start column 243 00:16:51,170 --> 00:16:57,710 width and height and angle to which on which we rotated our frame because we rotated to 90 degrees, 244 00:16:57,950 --> 00:16:59,000 we provided 90. 245 00:16:59,180 --> 00:17:04,790 And this simply updates the reference parameters that can be used later on when you want to reference 246 00:17:05,090 --> 00:17:08,900 on bear on the original frame B crop to get our frame of reference. 247 00:17:09,500 --> 00:17:16,250 Now finally, simply display all the intermediate stages and the final stage that we got by implementing 248 00:17:16,250 --> 00:17:17,810 the function of the extract background. 249 00:17:18,220 --> 00:17:24,710 This includes regional frame, Christmas frame column, ground replica burger model and main documents 250 00:17:24,710 --> 00:17:25,040 secret. 251 00:17:25,550 --> 00:17:30,770 This last two outputs are the final outputs of this particular function of extract background. 252 00:17:31,130 --> 00:17:37,580 So the background model will be used to use to get the foreground using the background subtraction and 253 00:17:37,580 --> 00:17:43,670 the mixed occupancy gear will simply be used for the next stage of mapping when we want to identify 254 00:17:43,670 --> 00:17:45,410 which region the caller controls. 255 00:17:46,010 --> 00:17:51,110 So now let's run this and see whether we were able to correctly implement the first step that is background 256 00:17:51,120 --> 00:17:55,640 extracted for the background subtraction task of local and state. 257 00:17:56,150 --> 00:17:59,690 So let's do this now before we can run this project. 258 00:17:59,780 --> 00:18:02,750 We still need to import it inside the main Salvador part. 259 00:18:03,170 --> 00:18:05,410 So let's go over the missile part. 260 00:18:06,380 --> 00:18:14,060 We're right at the start and import from our board localization import the class that is known as localizing. 261 00:18:14,750 --> 00:18:16,760 Then we simply create the object of this class. 262 00:18:18,110 --> 00:18:18,370 Four. 263 00:18:18,530 --> 00:18:25,190 And the object for the missile over and give it the same name as a class object that is bought. 264 00:18:25,190 --> 00:18:25,820 Localize it. 265 00:18:27,410 --> 00:18:30,920 We initialize this to the localize that object. 266 00:18:31,820 --> 00:18:38,090 So once we have the object created for this missile class, we can use this object inside the Mir solving 267 00:18:38,090 --> 00:18:38,570 function. 268 00:18:39,380 --> 00:18:40,040 We're writing. 269 00:18:41,270 --> 00:18:45,140 Silver that bought localized, localized sport. 270 00:18:45,650 --> 00:18:51,800 And that takes an argument of grand frame, which is basically sales or set view or the satellite view 271 00:18:51,800 --> 00:18:54,770 that was being provided to us from the top down camera. 272 00:18:54,800 --> 00:19:00,560 Looking on the means for the frame display, we need to create a new variable or placeholder that we 273 00:19:00,560 --> 00:19:01,640 name as frame display. 274 00:19:02,510 --> 00:19:05,240 And this is simply a copy of self thoughts that you. 275 00:19:09,180 --> 00:19:14,460 Now to provide this information, we can now simply build this project and see whether the full step 276 00:19:14,460 --> 00:19:17,160 of the localization states is working correctly or not. 277 00:19:18,030 --> 00:19:19,290 So we move inside the terminal. 278 00:19:21,400 --> 00:19:22,240 Will the project. 279 00:19:24,700 --> 00:19:25,900 Launch a simulation. 280 00:19:28,010 --> 00:19:32,510 So when this once the simulation starts up, we can now run the MIT solver. 281 00:19:33,800 --> 00:19:38,960 And as you can see, it has grown successfully and it has given me the five outputs. 282 00:19:39,140 --> 00:19:40,550 Let's see the original output. 283 00:19:40,640 --> 00:19:41,900 That was a satellite view. 284 00:19:42,110 --> 00:19:47,930 As you can see, we have the mess looked down upon from the satellite and we have the we have the board. 285 00:19:47,940 --> 00:19:48,350 Right. 286 00:19:48,350 --> 00:19:50,600 And this miss entry point of the miss. 287 00:19:51,290 --> 00:19:54,470 Then for the first output that we got was the region of interest lost. 288 00:19:54,950 --> 00:20:00,290 Now these models are all of the region of interest that represent inside the frame that included the 289 00:20:00,290 --> 00:20:01,820 Miss Walls and the car. 290 00:20:02,270 --> 00:20:06,950 Now, what we did was simply extracted the car mask and simply removed it. 291 00:20:07,980 --> 00:20:09,120 From our mainframe. 292 00:20:09,240 --> 00:20:14,880 And what we got as an output was this frame car removed, which included the frame, but the car was 293 00:20:14,890 --> 00:20:16,000 removed altogether. 294 00:20:16,290 --> 00:20:18,900 But this left an empty space right here. 295 00:20:18,990 --> 00:20:21,360 So we fill that space with a ground replica. 296 00:20:22,420 --> 00:20:27,850 So the ground replica was simply of the same color of the as of the bag known for the mace. 297 00:20:28,540 --> 00:20:31,210 So we simply filled the space that was left empty. 298 00:20:31,210 --> 00:20:36,820 And the final output, that record was the background model which included the mace inside the frame. 299 00:20:37,090 --> 00:20:39,940 But the foreground that was the car was simply removed. 300 00:20:41,090 --> 00:20:45,260 Once we have the background model, the first output, we can now move on to the second output that 301 00:20:45,300 --> 00:20:49,520 will see on this particular step of the block localization states. 302 00:20:49,970 --> 00:20:52,160 The second output was the occupancy grade. 303 00:20:52,430 --> 00:20:57,500 And as you can see, not only the maize has been dropped out to get the frame of reference, it has 304 00:20:57,500 --> 00:21:00,710 also been rotated to get the maize entry right on top. 305 00:21:01,070 --> 00:21:04,610 So this will be our frame of reference on which the board will be localised. 306 00:21:04,940 --> 00:21:08,000 So the port is on right here on the maize entry. 307 00:21:08,240 --> 00:21:13,580 It is basically on the top right position and can be interpreted as the frame. 308 00:21:13,970 --> 00:21:19,820 The columns of the frame are a few rows, a few columns before the next column of this particular frame. 309 00:21:20,120 --> 00:21:21,600 So this is our frame of reference. 310 00:21:21,620 --> 00:21:25,850 The second output of the extract background function for our local and state. 311 00:21:26,300 --> 00:21:32,450 Now let's implement the second step to use the to use this full background model that we have just expected 312 00:21:32,870 --> 00:21:35,010 to extract the foreground that is the crown. 313 00:21:35,030 --> 00:21:37,250 In the next few frames till then. 314 00:21:37,520 --> 00:21:38,120 Thank you. 315 00:21:38,180 --> 00:21:38,750 And by.