Research On Key Technologies Of Scanning Panoramic Imaging And Early Warning System

Compared with traditional radar active detection equipment,visible light passive detection equipment does not need to actively send detection signals when detecting the enemy,making it naturally capable of resisting the "four major threats" encountered in traditional detection.In the passive detection system,the image-based panoramic situation observation and early warning system can obtain panoramic images around the protection area,detect,track and warn the targets in the image,and can deal with asymmetric threats.In addition,video surveillance,robot vision,and vehicle peripheral vision systems in the civilian field also have the need to obtain panoramic images.At present,there are three main ways to obtain panoramic images:(1)rotational scanning panoramic image acquisition method;(2)catadioptric panoramic imaging method;(3)special structural form to realize panoramic imaging.This paper introduces a system that obtains the panoramic image of the panoramic view through the weekly scan and finally realizes the early warning by analyzing the image,and introduces the key algorithms used in the system according to the process of image acquisition and processing.The main work and innovations of this paper are summarized as follows:(1)This paper introduces the acquisition and display technology of the weekly scanning panoramic image,and according to the imaging characteristics of the weekly scanning early warning system,an improved phase correlation image registration algorithm is proposed,it improves the operation efficiency by selecting the local area of the image to be registered for registration,and reduces the time consumption to29.7% of the traditional algorithm;(2)The most popular moving target detection algorithms are analyzed and compared.When the Vi Be algorithm is used to detect moving objects,a method of rejudging the foreground points with the Otsu threshold of the current frame image is proposed,which solves the "Ghost" problem of the Vi Be algorithm,and the flood filling algorithm is used to solve the hole problem in the detection result of the Vi Be algorithm;(3)The background knowledge of tracking algorithms and the basic theory of correlation filtering algorithms are introduced,aiming at the target tracking problem of the weekly scanning early warning system,an improved KCF algorithm based on scale estimation and improved template update strategy is proposed,it improves the target tracking success rate by 10.6% compared with the traditional KCF algorithm,and the average tracking rate is 4.4 times that of the SAMF algorithm and 2 times that of the DSST algorithm;(4)Experimental simulations of the proposed algorithm are carried out,and the effectiveness and applicability of the local phase correlation algorithm,the improved Vi Be algorithm and the improved KCF algorithm are proved.A prototype of a weekly scanning panoramic system is built,and the image stitching algorithm proposed in this paper is integrated into the system and runs stably.In addition,the YOLO object detection algorithm is integrated into the system as its auxiliary function.