论文部分内容阅读
为了使算法更适应实际系统的要求，并能在实时系统中应用，研制了能自动给4个模板与实时图像进行相关后的最佳匹配位置的相关器．通过软件试验，验证了多模板相关算法的实用性，并通过对多种类型目标的对比实验，指出了正8像素重叠的四模板能较好地兼顾跟踪小目标和大目标．
In order to adapt the algorithm to the requirements of the actual system and apply it in the real-time system, a correlator that can automatically give the best match position of the four templates to the real-time image is developed. Through the software experiment, the practicality of multi-template correlation algorithm is verified, and through the comparison experiment of many types of targets, it is pointed out that the four templates with positive 8-pixel overlap can better track small targets and large targets.