Research on obstacle avoidance method of unmanned helicopter based on binocular vision

The unmanned helicopter for the electric power inspection process is faced with the difficult problem of needing to make real-time obstacle avoidance decisions in a complex operating environment, and the use of the traditional single obstacle avoidance mode is unable to adapt to the increasingly complex operational needs. In this paper, the binocular vision system is used for target recognition and separation, and the parameters are optimized by calibration method; the BM3D algorithm is used for denoising the image set, as well as the semi-global stereo matching algorithm is used for matching the images; finally, the Gray Wolf Optimization Algorithm is used to optimize the obstacle response coefficients in the IFDS, and the smooth obstacle avoidance trajectory is obtained. The research foundation is laid for the subsequent introduction of more detailed obstacle information and more constraints on the unmanned helicopter's own dynamics.