Research On Stereo Vision Ranging System Assisting UAV For Accurate Hanging On Power Transmission Line

In order to hang the UAV(Unmanned Aerial Vehicle)accurately on the power line,it not only requires the ranging system to quickly and accurately sense the distance from the power line,but also requires the system to be small in size,light in weight and low in power consumption.The stereo vision ranging system can obtain abundant depth of scene only by matching stereo images of two cameras,and has the advantages of simple structure and easy miniaturization.However,the algorithm of stereo matching algorithm is large,and the depth graph frame rate of the algorithm with high matching accuracy is low,which makes it difficult to achieve a good balance between real-time performance and precision.In order to solve this problem,this article makes improvements from camera model to stereo matching algorithm,improving the precision of distance measurement.Also,it implements algorithm acceleration on FPGA(Field Programmable Gate Array),improving real-time performance of the distance information.In particular,based on the ideal linear model,this paper mainly considers the influence of the radial distortion of the lens on the projection of the space point in the imaging plane,and establishes a nonlinear camera model with high precision,and adopts the nonlinear optimization algorithm to calibrate the model.In the image algorithm,the images are processed by threshold transformation,which can separate the target from the background pixel.In view of this characteristic,this paper puts forward the adaptive weight stereo matching algorithm,the algorithm based on whether the pixel belongs to the field lines around the image corresponding areas pixels,assign different weights on cost aggregation,thus effectively improving the matching accuracy.In the generated scene depth map,this article adopts the hough transform to detect the straight edge of the field lines.On the basis of this step,distance information between the unmanned is further optimized,finally improving raging precision.In order to verify the effectiveness of the proposed algorithm,this paper makes a comparative experiment on the stereo vision ranging system.The improved and unimproved stereo matching algorithm is allocated to FPGA,and the power line ranging is in multiple positions.The experimental results show that the range error of the stereo vision ranging system with improved stereo matching algorithm is relatively small,which proves that the proposed adaptive weight stereo matching algorithm can improve the matching precision.Moreover,the output frequency of the system is 30 Hz.Therefore,the stereo vision ranging system designed in this paper can achieve a good balance between real-time performance and accuracy,and meet the requirements of assisting UAV for accurate hanging on power transmission line.