A detection and tracking algorithm for optical target

In the terminal guidance phase imaging target occupies most of the field and shape information is rich. The main requirement is to stable tracking one part of the target. The gray scale of original image is too intensive and target is difficult to detect, so a method of gray-scale image histogram equalization is proposed. To fast detect corners of the image, a method of two-dimension projection and slide convolution is proposed. At last to reach real-time and stable tracing, so a type-heart tracing method based on corner match is used. And basis the feature of this algorithm, a hardware platform based on FPGA and DSP is designed. FPGA is good at parallel pipeline and DSP is good at logical operation. The algorithm is used by one item, and the target tracing is favorable. The algorithm ensures this item is successful.