Research And Application Of Police UAV Image Mosaic And Panoramic Construction

Aiming at the current problems such as insufficient police forces,slow forensics,and poor environment,the UAV aerial system can assist police with the investigation,forensics and real-time command.However,when the police taking images of a case or accident scene by UAV,due to the UAV is limited by the flight altitude and the lens focus,the single aerial image has a small field of view and it is difficult to obtain the overall cognition of the investigation scene,so the UAV aerial images need to be stitched into a panoramic image with larger field of view,richer information,and higher resolution.The UAV aerial images usually have various shooting perspectives and irregular overlap owing to the influence of the investigation scene.Therefore,high-quality panoramic mosaic of police UAV images is the focus of current research.The main contents of this paper are as follows:The police UAV aerial images are susceptible to noise pollution and are prone to blurring and geometric distortion,so this paper studies three preprocessing techniques,including image denoising,image enhancement and image correction.The four feature extraction algorithms,SIFT,SURF,ORB,and BRISK,are compared and analyzed for the registration effect and time by experiments,and the ORB algorithm has good result,so an improved registration method based on ORB was proposed for the characteristics of aerial images.The improved algorithm adopts the idea of image block and the strategy of non-maximal suppression to make ORB feature points evenly distributed in image.The experimental results show that the improved algorithm raise the registration rate compared with the original algorithm,and the registration speed has greater advantages than other algorithms.Thus,the proposed method has good applicability to UAV aerial images with large size and rich information.In view of the police UAV aerial images used for the investigation and forensics,the quality of the stitched image is required to be high.Therefore,this paper proposes an improved criterion for searching the best seam-line based on the difference of color and geometric and saliency characteristics,combining Laplacian blending to realize the image fusion in overlapping regions.The improved fusion method is compared with other three algorithms through experiments.The results show that the quality of stitched image by using the proposed method is higher,the problem of ghosting is solved,and the detail information at the stitching seams is reserved better,so that make it better for police to analyze the situation.In order to better use the image information of the investigation area and accurately analyze the situation,this paper combines the improved registration algorithm and fusion algorithm based on the UAV platform to design a police UAV image mosaic system,which is of great significance for alleviating the insufficient police forces and improving the efficiency of investigation and forensics.