Dynamic Target Image Dsp-based Real-time Processing Technology Research

The technology of dynamic target tracking and image processing has been a great degree of development in the modern military. With the development of Micro-processing chip technology, the stability and the accuracy of tracking have also been greatly improved. Modern weapons are equipped with advanced fire control system, its dynamic performance indicator is the shooting test, and evaluation need to check the main pilot projects. In order to correctly weapons assessment and evaluation, development of tests to meet the needs of modern weapons and equipment, dynamic tracking and measuring system is very necessary.In this paper, the requirements of a project for the Changchun Institute of Optics designed a DSP-based real-time dynamic target tracking image processing system can be achieved by dynamic target identification and tracking. In order to meet the project real-time image acquisition and processing requirements, and complex algorithms in real-time image processing on the application, this paper uses DSP as the main computing components to high-capacity FPGA as a coprocessor, designed a high-performance image acquisition and processing card. The card can be achieved Fourier transform, point the domain to accelerate such operations, to achieve expedite the processing of the image. This system can accomplish redevelopment. Users of this system can develop the image processes program to satisfy the requirement. This image acquiring and processing card can fully satisfy the computing requirements by the actual experiment testing.This paper briefly describes the design and workflow software solutions, gives the solution to difficulties, analysis of a large number of tracking algorithms and specific implementations, and algorithms are given for different conditions and adapt accordingly improved methods, and its application to the actual hardware devices, so that these algorithms under different conditions can play their respective advantages, for dynamic target tracking in real-time processing.The system is divided into target acquisition and target tracking in two parts. In the process of target acquisition, Images are obtained through the camera, the go through a series of operation, image preprocessing, image segmentation, parameter extraction and target matching finally the target can be identified from the background. Capture stage will focus on a multi-frame difference multiplied by the moving target detection method detected target.. In the tracking stage:in several classic algorithms, such as point tracking, trajectory extrapolation fitting tracking, focus tracking and correlation tracking, the focus on tracking and correlation tracking of the center of gravity is proposed hierarchical search algorithm for fast correlation tracking. The background and conditions of the system can select the appropriate changes in the tracking mode, thereby completing the target location prediction.After testing the system to achieve focus and associated matching algorithms, tracking, better-performing, and has a high real-time, to meet the target range detection and location of the actual demand.