Stereo Vision Image Processing Strategy for Moving Object Detecting

Here, a TMS320C6416 DSK board is integrated with two CMOS color image sensors to construct a new stereovision platform instead of current PC based or multi-CPU combination structures for detecting moving objects. A novel moving object detecting strategy is proposed by combining the image temporal differencing and the sum of absolute difference (SAD) matching schemes. This system can detect and track any moving object without object color and shape limitations. The subtraction image processing technique is employed to detect the moving object first based on single image. Then, the edge and SAD stereo matching schemes are used to establish the coordinate relationship of left-right two images. Finally, the depth of a moving object can be estimated for 3D position calculation based on stereo geometry relationship. Experimental results are used to evaluate this system dynamic performance. Key-Words: Stereo vision, CMOS image sensor, and moving object detecting strategy.