Intelligent Security Surveillance Image Compression And Abnormal Behavior Identification

Information processing has become an important part in the development of scienceand technology. The monitoring system can be seen everywhere in today’s society,including in the bank, transportation, office buildings, private residences and other places.They have become one of the indispensable equipment to protect the safety of people andproperty. The main purpose of monitoring system is a real-time monitoring of thedetection range, recording what is happening, and preserving. There are state-of-the-artmonitoring system with anomaly identification, voice alarm signal transmission,and so on.In Intelligent monitoring system, the camera needs to be real-time acquisition of eachframe image and save it on the hard disk, if unusual happens you need to send abnormalintelligence collected to the management staff and the acquisition of image compression isone of the important link. Due to the monitoring equipment collected image size is fixedand untreated. If we use it directly to transmit or save, it not only reduces the transmissionefficiency, but also take up a lot of memory space. Therefore, the collected imagecompression processing is very necessary. In this paper, we base on DCT transform imagecompression algorithm to process the collected image compression, and we not only retainmost of the information of the original image, but also ensure a higher compression ratio.Compared with other lossy compression algorithm this algorithm is relatively simple andeasy programming. In the monitoring system under the requirements of the quality of theimage is not particularly high environment, the use of the compression algorithm is enoughto meet the need.In the intelligent security monitoring systems, another important part is the exceptionacts of recognition. Because the abnormal situation always happen around the self-servicebank, ATM machine and other places, such as robbery, fighting and so on, people’spersonal safety and property safety pose a threat. Therefore, it is needed to identifyabnormal behavior of the part in video monitoring system. In the normal operation of the monitoring system, if you encounter a similar robbery, fighting behavior we use certainalgorithms to determine whether the act is abnormal. If it is determined as the exception,voice alarm and send a signal to the authorities, so as to resolve the exceptions and protectthe safety of persons and property of the people as soon as possible. In this paper theabnormal behavior recognition algorithm abandon the traditional complicated recognitionprocess. The traditional methods include track matching, which need a series ofcomplicated calculation, and need to build a large number of abnormal behavior library.This algorithm although accuracy is high, but the real time are not being met. In this paper,we use the centroid trajectory method, which the detection target is regarded as a point andthrough analysis of the trajectories points, it can be determined whether the behavior isabnormal. Proven that the algorithm is simple and feasible, and has some practical value.