8. Opencv application--Target tracking

Locating objects in consecutive frames of a video is called tracking. In OpenCV, you can use traditional target tracking algorithms (such as mean tracking, Kalman filter, etc.) or deep learning-based target trackers (such as MOSSE, CSRT, etc.) Perform target tracking. Deep learning object trackers generally perform better in accuracy and robustness

8.1. Use

Source code launch file path:/opt/ros/noetic/share/opencv_apps/launch

Step 1: Start the camera

xxxxxxxxxx
roslaunch astra_visual opencv_apps.launch img_flip:=false

• img_flip parameter: whether the image needs to be flipped horizontally, the default is false.

The [usb_cam-test.launch] file opens the [web_video_server] node by default, and you can directly use the [IP:8080] web page to view images in real time.

Step 2: Start the corner detection function of Opencv_apps

xxxxxxxxxx
roslaunch opencv_apps camshift.launch # Target tracking algorithm

Each functional case will have a parameter [debug_view], Boolean type, whether to use Opencv to display images, which is displayed by default.

If no display is required, set it to [False], for example

xxxxxxxxxx
roslaunch opencv_apps contour_moments.launch debug_view:=False

However, after starting in this way, some cases cannot be displayed in other ways, because in the source code, some [debug_view] is set to [False], which will turn off image processing.

8.2. Display method

• rqt_image_view

Enter the following command to select the corresponding topic

xxxxxxxxxx
rqt_image_view

• opencv

The system displays it by default, no need to do anything.

• Web viewing

(Same as under LAN) Enter IP+port in the browser, for example:

xxxxxxxxxx
192.168.2.116:8080

For specific IP, use your current virtual machine IP.

8.3. Effect display

You can see an adjustable window appear on the screen, followed by a red frame.