Research on image preprocessing and target detection of vehicle-machine cooperative system

In the present cooperative competition of robot, the robot needs to use its own camera module to collect real-time images and identify the enemy robot according to the collected images. However, due to the interference of some external factors in the arena, the contrast and clarity of the images collected by the robot are poor, which is not conducive to the accurate recognition of the robot. Therefore, this paper will preprocess the images collected by the robot, and reduce the interference caused by off-site factors through image filtering, image enhancement, and grayscale processing. Then use the method based on HSV color detection to achieve accurate identification of enemy robots. Finally, the accurate coordinates of the enemy robot can be obtained by using the target contour method to meet the accuracy requirements. Through the above methods, the robot can quickly and accurately obtain the position and shape of the enemy robot.