Moving Target Detecting And Tracking Based On DSP

Automatic tracking of moving target is a technology which can detect, pick-up, identify and follow a target. The technology is the core of computer machine vision, which can be widely used in many fields, such as in military affairs,traffic,biology,medicine and so on.Digital Signal Processors (DSP) provide strong groundwork for the implementation of real-time image processing software and hardware. Many problems in automatic tracking of moving target algorithm: plentiful operations, complicated implementation in hardware and the high cost are solved.Based on a series of experiments on TI ICETEK-DM642, the characteristics of some algorithms are analyzed and compared in this dissertation: the half-baked trait of picking up target through frame difference, the limited trait of using situation through background subtraction, the complexity of picking-up target center of gravity on the next step, and color detecting algorithm having to search at the whole image. Based on the combination of figure detecting and color detecting algorithm, a new target detecting algorithm is proposed in this paper, in which a target area is locked quickly through frame difference algorithm, and the target is searched through detecting algorithm at this target area in YUV color space. The experiments on TI ICETEK-DM642 prove that a red ball target is gained soon and effectively, rarely not suffering from the beam.On the basis of the Binocular Vision Model, the three-dimensional coordinates are gained through advanced image processing including target detecting, the picking-up of the target center of gravity, and the feature point matching. The motion parameters are calculated to drive the direct current motor. Based on the hardware of ICETEK-DM642 and ICETEK-F2812, a new kind of high-speed digital image detecting and tracking system is built, and the tracking experiments are primarily carried out on this system, preliminary results are reached.