Video Target Tracking System Reusable Framework For Research And Design

The thesis is focused on object-based algorithms designing and reusable framework of image processing. As a widely used technology, video analysis is task-dependent, different algorithm is chosen in allusion to different applications. Usually, people add some special assistant operations into the algorithm to test and analysis. Researching the intrinsic connections among the algorithm and the object-oriented programming is a significative thing. Then, with deeper understanding of them, we hope to make programming more easily to algorithm designers.In the aspect of relationship between algorithms and objects, we begin with the adaptive color and texture segmentation for tracking moving objects, then analyze the weakness of original algorithm flow and give an improved method to its structure. Afterwards, the thesis inducts the object-based thinking into algorithm design. Experiments result show it works more flexibly for algorithm design and test.Next, we explore the programming design in video processing system, and expand it to use design patterns to resolve problems. In the end, the thesis give an integrate design of image processing system, and a relatively integrate reusable framework based on it.