Research On Super-resolution Reconstruction Methon Of GM-APD Lidar Range Image Based On Intensity Image Guidance

Geiger mode avalanche photodiode(GM-APD)lidar has fast imaging speed and strong detection ability,and can be applied in many hotspots.However,the current range image obtained by this type of lidar has poor image quality and low spatial resolution.Therefore,it is necessary to perform range image analysis.Super-resolution reconstruction.In response to this problem,this research explores the image-guided range profile reconstruction algorithm,and analyzes several shortcomings of the algorithm:pixels deviate from the true value,edge regions are blurred and unclear,and the details of non-edge regions are poor,and then address its problems.An improved intensity image-guided super-resolution reconstruction algorithm is proposed,which improves the image quality,range resolution and spatial resolution of the GM-APD lidar range image.First,this research explores the principle of the image-guided range profile reconstruction model,and derives the detailed formula for solving the optimization equation in the algorithm.The simulation data and real data are respectively used as the input of the super-resolution algorithm.The results show that,Due to the lack of constraints in the data fidelity items,there are too many neighborhood-related calculation processes in the constraint items,and the lack of edge constraints,resulting in the reconstruction results deviating from the real situation,and the edges are blurred,and the details of non-edge areas are poor.Secondly,in view of the problem of unclear edges and poor detail information in the reconstruction results,this research proposes image registration methods,image preprocessing methods,gradient constraints and edge constraints;for the problem that the reconstructed range image deviates from the real situation,this research An optimized data fidelity item is proposed.Then use simulation data to verify the optimization items proposed above.The results show that,compared with other algorithms,the optimized and improved algorithm proposed in this study has clearer edges of the reconstruction results,richer detailed information,and optimal indicators.Finally,based on the lidar imaging experimental system,a verification experiment for image quality,range resolution,and spatial resolution was designed.Using real data obtained in outdoor experiments,the algorithm proposed in this study is in terms of image quality,range resolution,and spatial resolution.The impact of the rate is verified,including the impact of algorithm improvement,the impact of image registration and the impact of preprocessing,and the reconstruction results are evaluated.The results show that the distance value of the reconstruction result of the improved algorithm is close to the real situation,and its edge is sharp and real,with rich and accurate detail information,and the distance resolution of the reconstruction result is 19 cm,which is better than the original algorithm processing result and the original distance.The image is improved by 1cm and 2cm respectively;the spatial resolution of the reconstructed result is 4mm,which is 3mm higher than the original range image.