Weak Neural Network-based Moving Target Detection And Tracking

Video-based motion analysis aims at detecting, tracking and identifying moving objects, and more generally, understanding objects behaviors through analysis and processing image sequences with moving objects.The detection and tracking of the moving objects is one of the most important branches in the computer vision,which combines advanced technologies and research achievements in image processing,pattern recognition,artificial intelligence,automatic control and other relative fields.It has broadly applied in video surveillance,robots navigation,video transmission,video retrieval,medical image analysis,Meteorological analysis and other fields,so this subject has important theoretical significance and wide practical value.The automatic target detecting and tracking which is based on BP(Back Propagation) nets is studied in this paper. On the basis of summarizing and analyzing actuality research and algorithms both here and aboard, an algorithm of moving objects detecting in the outdoor scenes is proposed. And this paper offers a kind of programming realization of using BP network to detect and track the moving objects.Further,through the realization course of the recognition algorithm,this paper analyses the concrete thinking and method of using BP network in pattern-recognition application of picture objects. The main work is as follows:1, Firstly,video images are converted to the binary through the pretreating process ,then confirming the objects area and tracking them. On the basis of pretreatment,features of moving object are extracted.2, Adopt the method of'neighbour to pels'to found a relation between import and outport.That is using 3 multiply 3 template slipping on the images,and sampling the image by the center of the spot.It not only reduce the dimenion of the picture,but also improve the speed of the training.3, BP network classifier based on small sample is built.Labeled samples are used to train the classifier.The problem of non-linear object classification can be solved better.The trained classifier can be used to detect and track unknown object samples. Simulation results show that this method can recognize the moving objects,and restrain the background noise efficiently.In conclusion,this method is concise and efficient.It can provide good filtering results and robust adaptability to image targets with clutter background.