Research on Image Navigation Algorithms for UAV

Abstract

UAV (unmanned aerial vehicle) is used in agriculture filed to inspect water conservancy facilities, it not only can greatly reduce labor, but also effectively save relevant costs. The choice of UAV navigation mode is the key problem to achieve patrol of water conservancy facilities. In order to solve this problem, a new method of UAV image navigation is proposed which is different from traditional image navigation, and experimental research is carried out. Firstly, the acquired image is corrected and preprocessed to reduce image noise by perspective transformation. Then, the target information is extracted by threshold segmentation and ROI (region of interest) method. Finally, the attitude information of UAV is calculated and displayed in the original image through OpenCV (Open Source Computer Vision Library) library function. Based on the calculation results, the attitude of UAV can be adjusted by flight controller.