Infrared Video For Moving Target Detection And Tracking Algorithm And Its Dsp Implementation

Infrared Video Target Detection and Tracking System plays a great role in the field of digital image processing, and is widely used in civil and military affairs. The moving target detection techniques on the region of interesting (ROI) discussed in this dissertation can be also applied in image compression, information retrieval and interaction control by ROI.According to the real-time feature of this tracking system, it is divided into four states: search, capture, track and memory, and the states transition is fulfilled in the condition of the excitation of input and transition condition. According to system requirement, we design hardware and software system as a whole, and also discuss all modular functions of hardware system and software system architecture design at details. Through establishing the equation of moving target detection and tracking system, we find out the methods to solve the equation, and from the methods we come to discuss several crucial techniques of the implementation of system, and also according to the feature of the infrared video image sequence which we are processing and analyzing, we bring forward moving target detection and tracking technology based on background moving estimation and compensation.And then we go further into the implementation plan and technological path of the four fundamental states of this system, implement and analyze these four state model blocks on the PC simulation system and C6x system, also present the results.At last we discuss the DSP implementation of this system and analyze the data flow of DSP implementation. We have done the related experiments on the video sequences, which were provided by the relevant institute. The results of our experiments show that the system designed is correct, the key technologies have been implemented successfully in practical system and the design of system hardware and software is successful too.