Problems encountered in installing Yolo3 target detection module in Autoware

"The predecessors planted the trees, and the later generations enjoyed the shade!" The problems encountered in the experiment were summarized on the basis of other netizens, so thanks to the bloggers who provided help!

Written in the front: Since I am doing target detection based on the fusion of camera and lidar information, I use the Autoware framework for experiments, and install the Autoware open source software framework in the early stage. The specific installation steps are listed by many netizens in the blog.To repeat, the installation tutorial is recommended: https://blog.csdn.net/yourgreatfather/article/details/86504547

I installed it successfully under the tutorial of this netizen.

Problem description: Since Autoware->Computing integrates YOLO3 related packages, as shown below.

So to install YOLO3 in Autoware, you only need to create a new data folder in the corresponding vision_darknet_detect/darknet/ directory, and then download the weight model trained by YOLO to the data folder.The command is as follows:

cd datawget https://pjreddie.com/media/files/yolov3.weightswget https://pjreddie.com/media/files/yolov2.weights

Run YOLO3 detection:

roslaunch vision_darknet_detect vision_yolo3_detect.launch

There is a problem after running: Error:cannot launch type of node[vision_darknet_detect/]...

Problem solution 1: Reference: https://blog.csdn.net/banzhuan133/article/details/78177676

Solution 2: There may be a package vision_darknet_detect with the same name, use the following command:

catkin_find --without-underlays --libexec --share vision_darknet_detect

Check if there is a package with the same name. After the query, there are two packages with the same name in different paths. Just delete the package under the path where YOLO3 is not installed.

The two vision_darknet_detect paths in my computer are: /home/yyl/Autoware-master/ros/devel/share/vision_darknet_detect and /home/yyl/Autoware-master/ros/src/computing/perception/detection/vision_detector/packages/vision_darknet_detect

After I delete vision_darknet_detect in the first path, I can run YOLO3, and the test results under Autoware->Computing can also display the real-time detected objects.