1 00:00:01,070 --> 00:00:02,840 Welcome to the third project. 2 00:00:03,050 --> 00:00:08,690 In this project, you will create your own Ross service interface and implement it into your code using 3 00:00:08,690 --> 00:00:10,970 what you learned from the previous lecture. 4 00:00:11,390 --> 00:00:17,330 For this project, you will make two nodes one node which acts like a client, which sends a request 5 00:00:17,330 --> 00:00:23,360 of an angle to a robot to turn a camera and another node which takes in the request and sends back a 6 00:00:23,360 --> 00:00:29,510 response, which will be an image the robot takes after moving the camera to the angle specified. 7 00:00:29,960 --> 00:00:34,880 Now, since we do not have a real robot and haven't learned about making robots in simulation yet, 8 00:00:34,880 --> 00:00:41,060 we're just going to have some pre taken images which are named by an angle and return one of these images 9 00:00:41,060 --> 00:00:42,380 based on the request. 10 00:00:42,830 --> 00:00:46,730 These images are available in the resources folder or you can create your own. 11 00:00:47,490 --> 00:00:52,410 It may be helpful to spend some time to think about what message types would be most ideal for our request 12 00:00:52,410 --> 00:00:53,550 and response. 13 00:00:53,760 --> 00:00:56,850 So let's break down what this project will look like. 14 00:00:58,220 --> 00:01:03,860 For this project, we will create two nodes a service client node and a service server node. 15 00:01:04,160 --> 00:01:09,680 The service client node will send a service request of how many degrees to turn the camera to the service 16 00:01:09,680 --> 00:01:10,700 server node. 17 00:01:11,780 --> 00:01:16,160 The server will then return the pre taken image data back to the client. 18 00:01:17,270 --> 00:01:21,080 Then as an added feature, see if you can make your service client. 19 00:01:21,080 --> 00:01:23,480 Display the image for the user to see. 20 00:01:25,400 --> 00:01:30,590 In order to complete this task, I will use the OpenCV Library, which is a commonly used computer vision 21 00:01:30,590 --> 00:01:31,280 library. 22 00:01:31,310 --> 00:01:36,800 I know we haven't explicitly gone over OpenCV in this course, so I'll quickly go over some useful functionality 23 00:01:36,800 --> 00:01:38,330 relevant to this project. 24 00:01:38,600 --> 00:01:40,940 In order to use OpenCV, you must install it. 25 00:01:40,970 --> 00:01:47,390 We will also utilize the complementary CV bridge package to help us convert between ROS and OpenCV data 26 00:01:47,390 --> 00:01:49,340 types when working with images. 27 00:01:49,640 --> 00:01:54,620 We need to also make sure we include dependencies for OpenCV in our package XML file as well as our 28 00:01:54,650 --> 00:01:56,240 C list text file. 29 00:01:57,650 --> 00:02:02,150 To import OpenCV in your python code simply call import CV to. 30 00:02:02,300 --> 00:02:07,790 Then we can also import the CV bridge object from the CV bridge ROS package. 31 00:02:08,210 --> 00:02:12,380 The common utility when using OpenCV is to open an image file on your computer. 32 00:02:12,410 --> 00:02:18,890 We can do this with the image read function shorthanded as I am read, followed by a string with the 33 00:02:18,890 --> 00:02:20,000 file location. 34 00:02:20,410 --> 00:02:26,030 You want to display an image, you can use the image show function which will open a window with the 35 00:02:26,030 --> 00:02:28,130 corresponding image you want to display. 36 00:02:28,430 --> 00:02:33,500 Then we can use the weight key function with a value of zero to signal that the window should stay open 37 00:02:33,500 --> 00:02:36,860 until we close it with the enter key on our keyboard. 38 00:02:37,220 --> 00:02:41,810 Then lastly, you can also have your code ensure all OpenCV display windows are closed when finish running 39 00:02:41,810 --> 00:02:44,630 your script by calling destroy all windows. 40 00:02:45,020 --> 00:02:47,580 So that's just about covers all you need to know about OpenCV. 41 00:02:47,600 --> 00:02:49,940 So now let's jump into CV Bridge. 42 00:02:50,030 --> 00:02:54,890 Once you've written the image, you may want to convert it to a ROS image message from the sensor messages 43 00:02:54,890 --> 00:02:56,030 ROS package. 44 00:02:56,090 --> 00:03:03,530 You can do this by calling the CV to to image a message method of the CV bridge object, then to convert 45 00:03:03,530 --> 00:03:09,500 a ROS image message back to an open CV data type, you can use the image of message two CV two method. 46 00:03:10,230 --> 00:03:15,000 Also a helpful reminder on how to use the image interface type in your custom service. 47 00:03:15,030 --> 00:03:20,670 Here's an example of how to use types from a different Ross package and how to include these ROS packages 48 00:03:20,670 --> 00:03:22,440 as dependencies within your CE. 49 00:03:22,470 --> 00:03:24,030 Make list text file. 50 00:03:27,260 --> 00:03:27,680 All right. 51 00:03:27,680 --> 00:03:32,710 With that OpenCV prep out of the way, it's time for you to try this project out on your own. 52 00:03:32,720 --> 00:03:38,660 So go ahead and download the sample images from the resources section of this lecture. 53 00:03:38,900 --> 00:03:45,440 Then make your custom service, which will allow you to request a degree for the camera to turn, which 54 00:03:45,440 --> 00:03:47,180 responds with an image. 55 00:03:47,420 --> 00:03:53,600 Then, with the help of OpenCV, create a service server which will take the request and return an image 56 00:03:53,600 --> 00:03:59,150 from the sample images, as well as a service client which will be used to make the request. 57 00:03:59,150 --> 00:04:04,430 And as an optional challenge, display the image returned in an open CV window. 58 00:04:04,730 --> 00:04:10,040 With that, pause the video and give yourself an hour or so to try and complete as much of this project 59 00:04:10,040 --> 00:04:11,840 as possible on your own. 60 00:04:20,180 --> 00:04:20,560 All right. 61 00:04:20,570 --> 00:04:21,350 How'd it go? 62 00:04:21,440 --> 00:04:27,110 I'm sure this was a hard project being that you had to use OpenCV, and we have not explicitly covered 63 00:04:27,110 --> 00:04:31,490 it in this course, but this is exactly how it feels to be a robotics developer. 64 00:04:31,760 --> 00:04:36,590 You may have also been able to achieve the same result with a different framework or your own custom 65 00:04:36,590 --> 00:04:37,400 solution. 66 00:04:37,400 --> 00:04:42,620 But as a developer, you always need to balance how to incorporate advanced functionality into your 67 00:04:42,620 --> 00:04:44,900 robots as quickly as possible. 68 00:04:45,200 --> 00:04:48,560 I will go ahead and walk through how I would approach this project. 69 00:04:50,750 --> 00:04:56,360 In order to do this project, make sure you have downloaded the images from the resources section here. 70 00:04:56,360 --> 00:05:01,940 I've gone ahead and downloaded it and unzipped it in my downloads folder, so I'll just go into my package 71 00:05:01,940 --> 00:05:02,840 directory. 72 00:05:09,070 --> 00:05:10,990 And copy the images into their. 73 00:05:15,350 --> 00:05:15,680 All right. 74 00:05:15,680 --> 00:05:16,950 Let's go ahead and close that. 75 00:05:16,970 --> 00:05:22,730 And the other thing to note is I will be using the OpenCV Python module to utilize some of its image 76 00:05:22,730 --> 00:05:27,040 reading, processing and display functions in order to install it. 77 00:05:27,050 --> 00:05:34,910 You can open up a terminal and update your app repository and then install lib opencv dash dev and python 78 00:05:34,910 --> 00:05:36,830 three dash OpenCV. 79 00:05:44,640 --> 00:05:47,700 And so in this case, I've already gone ahead and installed them ahead of time. 80 00:05:47,700 --> 00:05:50,490 But be sure to go through that installation process. 81 00:05:50,610 --> 00:05:57,870 We can double check that this installation worked by starting Python three in terminal and importing 82 00:05:57,900 --> 00:05:59,130 CV two. 83 00:06:01,610 --> 00:06:08,740 And then we can go ahead and run the CV to DOT and then to underscores version and to underscores again, 84 00:06:08,750 --> 00:06:11,930 note that if you have a similar or newer version, you should be fine. 85 00:06:11,930 --> 00:06:17,030 But here we can confirm I'm able to import CV two in Python and here we can just reference what version 86 00:06:17,030 --> 00:06:19,580 number we have so I can go ahead and exit this. 87 00:06:21,610 --> 00:06:22,540 And clear the screen. 88 00:06:23,140 --> 00:06:28,900 OC First thing I'm going to do is open up V's code and create a new custom SVG file. 89 00:06:28,930 --> 00:06:32,680 I'll call this RV file Turn camera, RV. 90 00:06:32,830 --> 00:06:37,660 For this project, I want to take in a number in degrees and return an image. 91 00:06:37,660 --> 00:06:44,140 So I will use float 32 for my request and call the attribute degree turn then separate the request and 92 00:06:44,140 --> 00:06:50,470 response section with three dashes and then for the response I will use the image message type from 93 00:06:50,470 --> 00:06:52,720 the sensor messages library. 94 00:06:52,750 --> 00:06:55,330 I will just call this response attribute image. 95 00:06:55,840 --> 00:06:56,230 All right. 96 00:06:56,230 --> 00:07:02,380 With that, I can save this file and head over to my make lists, text file, and we can go ahead and 97 00:07:02,380 --> 00:07:06,160 add it to our ROS Idle generate interfaces Command. 98 00:07:08,330 --> 00:07:14,150 Note that I am also using the image message from Sensor Messages library, so I need to be sure it'll 99 00:07:14,330 --> 00:07:20,120 make that I also need that package and that it's a dependency for these ROS idle interfaces. 100 00:07:21,920 --> 00:07:29,120 So I can do that with the dependencies keyword and followed by the name of the particular package that 101 00:07:29,120 --> 00:07:30,050 it's depending on. 102 00:07:30,200 --> 00:07:36,410 With that I can go ahead and save the file and we can head over and build the workspace. 103 00:07:39,180 --> 00:07:41,070 All right, so I can go ahead and close this. 104 00:07:41,100 --> 00:07:46,050 Open up our terminal, make sure we're sourced to our install setup file of our workspace. 105 00:07:46,050 --> 00:07:51,840 And now I can run the Ross to interface list command, and I should be able to see our new turn camera 106 00:07:51,840 --> 00:07:53,910 service interface within my package. 107 00:07:54,730 --> 00:07:55,080 Great. 108 00:07:55,090 --> 00:08:01,960 Now I can use this interface in my code, so I'll go ahead and make my service server node within my 109 00:08:01,960 --> 00:08:03,070 scripts directory. 110 00:08:03,100 --> 00:08:04,840 I'll just call it turn camera server. 111 00:08:07,600 --> 00:08:13,720 As usual, I will just copy over the service server code that we created in the last lecture as a baseline. 112 00:08:17,980 --> 00:08:22,540 And I'll go ahead and change the imported service to turn camera. 113 00:08:26,680 --> 00:08:33,520 I'll also go ahead and change the class node service topic and callback function names to match the 114 00:08:33,520 --> 00:08:35,050 context of this project. 115 00:08:42,520 --> 00:08:48,880 And I'll go ahead and delete the current content we have in our callback function for our service. 116 00:08:49,780 --> 00:08:55,300 Now, remember, for this project I will have five sample images which I'm going to choose from. 117 00:08:55,300 --> 00:09:00,160 So I'll just make a note of these values up here in a variable called available angles. 118 00:09:01,570 --> 00:09:07,150 So when I receive a request, the first thing I will do is check this list to which number is closest 119 00:09:07,150 --> 00:09:09,160 to that of the number given by the user. 120 00:09:09,310 --> 00:09:14,410 To make this a bit more organized, I will create another function called Get Image, which will help 121 00:09:14,410 --> 00:09:18,910 with this processing by taking in the degree turn attribute of the request. 122 00:09:24,300 --> 00:09:26,280 Then I can define this function below. 123 00:09:36,970 --> 00:09:42,850 So I'll go ahead and calculate the closest angle by using the min python function and making a condensed 124 00:09:42,850 --> 00:09:43,970 lambda function. 125 00:09:43,990 --> 00:09:48,910 Basically, the min function will iterate the available angles list and I use the key argument to set 126 00:09:48,910 --> 00:09:52,060 a function to find the min value of the given list. 127 00:09:52,090 --> 00:09:56,740 In this case I take the absolute value of each available angle with the angle the user gives us. 128 00:09:59,260 --> 00:09:59,950 Don't worry too much. 129 00:09:59,950 --> 00:10:03,370 If you're not experienced with Lambda functions, you could have easily gotten the same results with 130 00:10:03,370 --> 00:10:04,740 a simple for loop. 131 00:10:04,750 --> 00:10:08,700 Then, just to make sure my code is working properly, I will print out the closest angle. 132 00:10:08,710 --> 00:10:14,470 Then I can recreate the file name of the image by combining the string of fied version of the closest 133 00:10:14,470 --> 00:10:19,420 angle in the available angles list and adding the dot PNG file extension. 134 00:10:19,540 --> 00:10:24,340 In the spirit of keeping the code tidy and organized, I'll make a new function called Read an image 135 00:10:24,340 --> 00:10:29,170 by file name which will be responsible for finding the image from within our images folder. 136 00:10:30,250 --> 00:10:32,170 So now I'll create this function below. 137 00:10:39,110 --> 00:10:44,390 Now, since we are working with reading and files, I will import the OSS Python module and use some 138 00:10:44,390 --> 00:10:45,410 of its functions. 139 00:10:45,920 --> 00:10:51,200 Now I'm going to want to find our image file dynamically rather than hardcoded in the file path in terms 140 00:10:51,200 --> 00:10:52,550 of where it is on my computer. 141 00:10:52,550 --> 00:10:59,330 So I will use the OSS path dur name function which will return the full file path of a file. 142 00:10:59,450 --> 00:11:03,410 In this case I'll just use the underscore underscore, file underscore. 143 00:11:03,410 --> 00:11:05,270 Underscore python variable. 144 00:11:05,300 --> 00:11:09,210 Note that this will not lead us to the script folder if we are using the ROS to run command. 145 00:11:09,230 --> 00:11:13,880 Instead, we will end up deep within the install folder of our workspace, which is where a copy of 146 00:11:13,880 --> 00:11:16,820 our script is made when we run an build. 147 00:11:17,300 --> 00:11:21,740 So I'll take this directory path string and index it for the install folder. 148 00:11:28,420 --> 00:11:33,670 Now I can crop the string using this index so we only get the file path of our workspace. 149 00:11:38,060 --> 00:11:39,170 Then I'll just take on that. 150 00:11:39,170 --> 00:11:44,690 We want to go into Source Udemy raster package and then the images folder. 151 00:11:44,810 --> 00:11:48,770 Make sure this path matches where you put the images within your package. 152 00:11:48,980 --> 00:11:54,080 Then lastly, I will just tag on the file name, which will be the corresponding number pad. 153 00:11:55,130 --> 00:11:55,610 Great. 154 00:11:55,610 --> 00:12:00,950 Now that I have the full file location of the image, I can go ahead and read it, in which I will do 155 00:12:00,950 --> 00:12:01,980 with OpenCV. 156 00:12:02,000 --> 00:12:04,220 So I'll go ahead and import at the top of the script. 157 00:12:06,410 --> 00:12:11,020 And since I'm here, I will actually go ahead and also import our CV bridge. 158 00:12:11,030 --> 00:12:12,290 And so I'm going to need that in a bit. 159 00:12:13,400 --> 00:12:19,310 This CV bridge is a ROS package, which you should have by default when we did our ROS installation 160 00:12:19,310 --> 00:12:23,330 and we can actually check this by heading over to our terminal and I'll click the screen and do Ros 161 00:12:23,330 --> 00:12:25,070 to package list. 162 00:12:32,060 --> 00:12:35,900 And here we indeed see the CV Bridge package. 163 00:12:38,630 --> 00:12:42,530 You can feel free to check out the OpenCV documentation if you want to learn more about the tools that 164 00:12:42,530 --> 00:12:43,880 are available and how to use them. 165 00:12:43,880 --> 00:12:49,310 For now, I will just head back down to my read in image function and we'll read in the image using 166 00:12:49,310 --> 00:12:53,840 the CV to I am read function which will take in the file location. 167 00:12:53,840 --> 00:12:56,000 Then we can go ahead and return that image. 168 00:12:56,750 --> 00:12:57,110 Great. 169 00:12:57,110 --> 00:13:04,370 So it returns the image to our get image function that we have here, which returns it to our original 170 00:13:04,370 --> 00:13:06,890 callback that we have here called Return Image. 171 00:13:07,430 --> 00:13:12,820 I now have to convert this image data into an actual ROS image message to load it into the response. 172 00:13:12,830 --> 00:13:18,170 Well, the CV bridge CV to to image message function can do that for us automatically. 173 00:13:19,550 --> 00:13:24,860 Then I can go ahead and take my response variable load the image attribute of it, since that is what 174 00:13:24,860 --> 00:13:30,020 I named it in my RV file and set it equal to this image message variable. 175 00:13:30,230 --> 00:13:32,870 Then I finish this function by returning the response. 176 00:13:33,650 --> 00:13:35,840 That should be it for our service server. 177 00:13:35,840 --> 00:13:40,220 So I'll go ahead and save the file and add it to our make list text file. 178 00:13:46,580 --> 00:13:53,840 Now go ahead and save the file and rebuild our workspace to include this new turn camera server Hi file. 179 00:13:57,340 --> 00:13:58,390 Now, let's go ahead and test this. 180 00:13:58,390 --> 00:14:01,270 I'm going to make sure that I'm source do my setup file. 181 00:14:07,070 --> 00:14:09,110 OC our server is running. 182 00:14:09,110 --> 00:14:11,780 So now I'll open a new terminal tab. 183 00:14:13,040 --> 00:14:15,020 And I'll source my workspace again. 184 00:14:17,640 --> 00:14:22,080 And I can use the Ross to service Terminal Command to interact with my server. 185 00:14:22,320 --> 00:14:27,330 In this case, I'll use Ross to service call to send a degree turn request of ten degrees. 186 00:14:27,330 --> 00:14:32,070 And again, I'm hitting the tab key here for auto completion to help me fill in these commands. 187 00:14:33,890 --> 00:14:37,000 OC We see a bunch of numbers flying across the terminal. 188 00:14:37,010 --> 00:14:41,500 This is actually the raw image data from the image our server sent back. 189 00:14:41,510 --> 00:14:48,670 And if we look above in our server terminal, we can see it calculated the closest image to be the 15th 190 00:14:48,680 --> 00:14:49,760 degree image. 191 00:14:49,760 --> 00:14:52,730 So it seems like our code is working just fine. 192 00:14:52,730 --> 00:14:58,220 So now I'll go ahead and make a client that can actually take the image data returned by the server 193 00:14:58,220 --> 00:15:00,640 and open a display window for us to see it. 194 00:15:00,650 --> 00:15:06,520 So I'll head back over to V's code and create a new script which I'll call Turn camera client Dot Pi. 195 00:15:08,610 --> 00:15:14,070 As usual, I'll go ahead and copy over the code that we created in the last lecture for our service 196 00:15:14,070 --> 00:15:16,290 client as a baseline. 197 00:15:18,260 --> 00:15:21,350 And I'll change the contents to match this project. 198 00:15:41,620 --> 00:15:41,920 All right. 199 00:15:41,920 --> 00:15:45,370 So we've gone ahead and created our turn camera client. 200 00:15:45,370 --> 00:15:53,500 And I still have this self request call, which takes our turn camera service and creates a request 201 00:15:53,500 --> 00:15:57,850 object, which we're going to be utilizing in this send request function. 202 00:15:57,970 --> 00:16:04,720 Now, instead of an integer, I can go ahead and convert this over to a float and we will have to change 203 00:16:04,720 --> 00:16:08,280 our request type to match what we have in our SRI file. 204 00:16:08,290 --> 00:16:14,440 So my request has the attribute degree turn, so I'll go ahead and match that naming convention. 205 00:16:17,290 --> 00:16:21,090 Then we'll go ahead and wait for the service server to be available. 206 00:16:21,100 --> 00:16:26,500 We have our future object, which we will make our asynchronous call with our degree turn that we have 207 00:16:26,500 --> 00:16:31,270 set up and spin until the future complete, until we return our result. 208 00:16:32,230 --> 00:16:35,080 So the only thing I've got to change now is if I head down over here. 209 00:16:35,920 --> 00:16:37,630 Change what our input is. 210 00:16:37,630 --> 00:16:44,320 So in this case we will say enter a number in degrees to turn the camera. 211 00:16:47,010 --> 00:16:52,230 Now, rather than just printing the result, we want to actually open a display window. 212 00:16:52,620 --> 00:16:57,180 So within my class, I'm going to create a new function called display image. 213 00:16:58,830 --> 00:17:03,750 And this is going to take in the image message that gets returned to us by our service server. 214 00:17:03,930 --> 00:17:08,790 Now, in order to complete this, I'm going to need my OpenCV and CV Bridge imports. 215 00:17:08,820 --> 00:17:11,970 So I'll go ahead and copy those from the previous file. 216 00:17:18,260 --> 00:17:23,870 And so now we're going to use this bridge again to convert the image message to the CV to format. 217 00:17:26,940 --> 00:17:32,070 Now I can use the open CV's show function to create an image display window. 218 00:17:32,100 --> 00:17:37,660 I'll name the window, turn camera image and pass it in the image variable I just made. 219 00:17:37,680 --> 00:17:43,260 Then I will use the CV to weight key function and set it to zero, which means that the window will 220 00:17:43,260 --> 00:17:45,690 stay open until I press the enter key. 221 00:17:45,720 --> 00:17:50,280 Then I can just make sure it closes all the open windows once we are done. 222 00:17:52,630 --> 00:17:55,540 And it seems my tabs are not configured properly. 223 00:17:55,540 --> 00:17:57,790 So just go ahead and fix this really quickly. 224 00:17:58,660 --> 00:18:00,040 And that takes care of that error. 225 00:18:01,250 --> 00:18:05,860 Then last thing I need to do is actually implement this function down in our main call. 226 00:18:05,870 --> 00:18:12,770 So instead of just doing our server returned, we will take our result and pass it to our display image 227 00:18:12,770 --> 00:18:13,430 function. 228 00:18:14,410 --> 00:18:14,790 Great. 229 00:18:14,800 --> 00:18:15,730 That's it for this node. 230 00:18:15,730 --> 00:18:22,090 Let's go ahead and save it and go over to our Quick List text file and make sure we install this program. 231 00:18:23,380 --> 00:18:24,820 So I'll save this. 232 00:18:24,880 --> 00:18:26,920 I'll go ahead and rebuild the workspace. 233 00:18:28,820 --> 00:18:29,090 All right. 234 00:18:29,160 --> 00:18:30,620 We can head back to our terminal. 235 00:18:31,630 --> 00:18:34,090 So our turn camera service server is still running. 236 00:18:34,090 --> 00:18:35,320 So go ahead and clear the screen. 237 00:18:36,250 --> 00:18:41,470 Make sure this workspace is source and run our turn camera client. 238 00:18:44,600 --> 00:18:44,930 Here. 239 00:18:44,930 --> 00:18:47,720 I get prompted how much I want to turn the camera angle. 240 00:18:47,750 --> 00:18:50,930 I'll just set it to ten like we did earlier. 241 00:18:50,930 --> 00:18:53,630 When we use the Ross to service, call command and terminal. 242 00:18:54,780 --> 00:18:58,980 Oh, and we get an error here saying attribute error turn and camera response object has no attribute 243 00:18:58,980 --> 00:19:03,300 encoding, so it seems like it's expecting an image, but it's getting our response object. 244 00:19:03,300 --> 00:19:07,650 So that means I forgot to edit our callback function in turn camera client. 245 00:19:08,160 --> 00:19:12,060 So here when we send the request, we get returned the entire result. 246 00:19:12,060 --> 00:19:15,630 So this encompasses the entire result object. 247 00:19:15,630 --> 00:19:22,680 So what I'm going to return instead is the image attribute of it, which is the attribute we named here. 248 00:19:22,920 --> 00:19:28,350 So now it's actually returning that image message instead of the response object. 249 00:19:28,350 --> 00:19:33,660 So if I go ahead and save that and we can go ahead and clear the screen and try rerunning this file. 250 00:19:35,880 --> 00:19:37,770 I'll try my ten again. 251 00:19:38,310 --> 00:19:44,640 And this time we get a display window which has the corresponding image for 15 degrees. 252 00:19:44,640 --> 00:19:47,520 Since that was the closest value we had to ten. 253 00:19:49,890 --> 00:19:54,750 Close this window, we can just go ahead and hit enter and we can go ahead and try this out with different 254 00:19:54,750 --> 00:19:55,260 pictures. 255 00:19:55,260 --> 00:19:58,500 So, for example, I could try -17. 256 00:19:59,790 --> 00:20:04,530 And they should give us the -15 image from our images folder. 257 00:20:05,840 --> 00:20:10,250 And you can keep trying this out for all the images we have in the folders corresponding to a certain 258 00:20:10,250 --> 00:20:10,850 degree. 259 00:20:12,510 --> 00:20:18,030 And with that, you have successfully completed this services project in which you implemented a Ross 260 00:20:18,030 --> 00:20:19,680 service to send a number. 261 00:20:19,680 --> 00:20:22,630 We wanted our fake robot to turn its camera to. 262 00:20:22,650 --> 00:20:27,990 And since we don't have a real camera, we instead just returned a pre taken image based on the request 263 00:20:27,990 --> 00:20:28,680 sent. 264 00:20:28,980 --> 00:20:35,850 You also learned how to incorporate the popular OpenCV framework into your raw code to process and view 265 00:20:35,850 --> 00:20:36,750 image data. 266 00:20:36,780 --> 00:20:41,850 So I hope this gives you a little taste as to how you can implement computer vision in your robotics 267 00:20:41,850 --> 00:20:42,810 applications.