Research On Image Recognition Algorithm Of Human Invasion Used In Security Monitoring

Automatic identification system of human invasion based on image processingwhich integrated multidisciplinary advanced technology overcome the weakness oftraditional intelligent security. It liberates labor and brings to people’s life safety,make efficient, convenient and reliable intelligent security to be possible. At the sametime, the image recognition algorithm of human body invasion used in securitymonitoring system which has the very high academic value and research significancehas become a hot topic in the field of computer vision research.This paper makes a study and analysis of two major problems involved in thealgorithm which is designed to realize automatic recognition of human invasion. Itcarried on the thorough analysis and put forward an effective solution, finallycombining theoretical research and practical application, make the image recognitionalgorithm is implemented in the security monitoring system. This article mainresearch content is as follows:1. This paper proposes a moving object segmentation method combined usingbackground subtraction and frame difference between. An illumination changeinterference removal method based on similarity histogram distribution is proposedfor may encounter illumination changes in the process of target segmentation. Themedian filter and morphological filtering method for interference removal are adoptedin this paper for possible noise signal as well as holes in target extraction. A shadowelimination method based on vertical projection is proposed in view of shadow causedby human body. Experimental results show that the method proposed in this paper caneffectively eliminate the influence of the corresponding interference which helps tosegment the moving targets more accurately.2. In terms of moving target identification, a method use the body geometrycharacteristics and characteristics of human face skin color matching is put forward.The geometrical characteristics of the human body target in video image is mainlystudied and a matching model based on geometric features of the human body is buildaccording to analysis of the characteristics that describing human geometricalparameters and the distribution of the parameter value. Then a method to describe thecharacteristics of human face skin color is studied and a recognition model based onthe characteristics of human face skin color is designed. In the end, a rough detectionbased on the geometric feature and fine detection based on skin color characteristics is implemented which realized fast and accurate human body target recognition.3. The image recognition algorithm of human invasion in softwareimplementation of security system is introduced, and then the development of imagerecognition algorithm of dynamic libraries under the framework of the Androidplatform is described in detail. Finally, the paper analyzes the application of thealgorithm in the actual system. The results show that this algorithm can achieveautomatic recognition of human body invasion under security monitoring system.