Study Of Keyhole Imaging Method For Objects Behind The Indoor Occlusion Area

In recent years,photoelectric imaging technology has been widely used in industry,medicine,biology and other fields.Especially in military and medical field,its application is more extensive.However,in the case where the target area is blocked or the field is restricted,the general imaging technique cannot obtain the target image.For example,in the field of military and counter-terrorism,when rescuing of indoor hostages,because the target is blocked by the wall,in the security or negotiation distance the conventional imaging system usually can not get the relevant information in time because of target occlusion or field limited,bringing difficulties for making rescue plans timely.It is necessary to make information acquisition and target imaging by means of unconventional detection,so combined with the feature of application scene,this paper proposes and studies a method of keyhole imaging detection based on array camera in order to solve these problems.In this paper,the method of keyhole imaging based on array camera can be used to image the indoor occlusion target through micro-keyhole.This method utilizes the array camera to shoot in the rear of the keyhole from many angles to obtain the multi angle video sequence,then obtains the target region video image through the video splicing technology,and finally achieves the goal of expanding the field of view.Based on the analysis of keyhole imaging theory,the method of extracting and splicing the target region is discussed.A method of aperture image extraction is proposed to obtain the effective target region after keyhole.By preprocessing each frame image,and using Canny operator and mathematical morphology method to locate the target region,the video image of the target area is obtained.After acquiring the image of the target sub region,we first use the SIFT algorithm to extract the feature and make the initial matching,then eliminate the mismatch.Secondly,the initial transformation matrix is solved by the projection transformation model and the Ransac algorithm.In order to realize the transformation matrix video stitching algorithm better,a new optimization algorithm of transformation matrix is proposed,and the image of the target region of greater field range is obtained by using the optimized transform matrix to splice the frame sequence of the subsequent video images.Based on the above analysis,the video stitching simulation experiment is carried out,and the feasibility of the transform matrix video stitching algorithm is validated by comparing the original video frame sequence with the video frame sequence image evaluation results.Finally,based on the simulation of the Keyhole imaging system,a 2x2 array imaging system is designed,and Condorcet image experiments are carried out in three different scenarios.Three experimental results show that the final splicing video is clear and natural,and the brightness transition is smooth.There is no obvious artificial stitching in the video.Thus,the applicability and stability of the system are verified.