Detection Of Human Abnormal Behavior In The Security And Defense Of The Ship

Ships is a national maritime combat forces, mainly using for the safeguard of thecountry’s maritime rights and interests, to defend the territorial seas not be violated The alertof China’s naval security usually rely more ground crew to protection in the non-wartime state,which is difficult to ensure the real-time monitoring of the security in the combat readiness ofthe case. According to some of the problems that exist in the actual ship security, this paperdid some research on ship security about single target abnormal behavior detectiontechnology.This paper firstly make the study and contrast of some methods of human detection bymultiple video from actual background of the Chinese People’s Liberation Army Navy’s104-ship shot, under the case of a fixed camera and PTZ. By analyzing the test results in thecase of priority detection accuracy and interference suppression, the paper selected thecodebook model under the YUV space as the background modeling algorithm. Secondly, forthe goal of breaking the code of the test results and lost, the text prospects background ratioand updates of time restrictions in the background update，to solve negative due to real-timeupdates of the codebook background model impact, better to overcome the impact of light,noise and shadows, and eventually made good detection results.According to the analysis of the codebook test results, the paper achieves a stable targettracking by using the combination of Camshift and codebook detection target trackingalgorithm, based on which by using Hu invariant moments to extract the characteristics of thehuman body, to establish the normal behavior template Library. Take average filtering processof the real-time detection of human characteristics. Abnormal behavior of human isdifferentiated by matching template with the normal behavior and combining with the aspectratio of the rate of change.Next, the paper do experimental research on the shaking and the waves of ships in theactual sea conditions noise: Solving the impact of ships shaking by using based on the changeof target centroid and codebook modeling combined; by maring deck shoot camera screenfixed coordinates, to determine whether the occurrence of abnormal behavior deckcoordinates intersect, thereby inhibiting wave noise interference.Experimental results show that: the algorithm using in simulate shooting video of thehuman abnormal behavior detection can be achieved very good results in the case of shipsmoored at the pier; when ships anchor in the sea environment, expected effect can be achieved through algorithm optimization, but there are still some gaps away from practicalapplication.