从流形群运运目标的形状特征及其本质属性出发，提出了多种适合计算机自动识别与跟踪流形群运运目标的匹配模板的构建技术。充分顾及流形群运动目标的区域整体描述、几何形状特征、不变矩特性，以及流形群目标的本质属性(纹理特征)，通过最大欧几里得贴近度的择近原则，实现模板匹配，完成目标识别；通过交替更新匹配模板元素参量及交叉匹配算法，实现目标跟踪。应用这种算法，对1998年6月至8月的青藏高原上空中尺度对流系统(MCS)进行了识别与跟踪实验。实验结果表明，这种技术较适合计算机自动识别和跟踪类似MCS的流形群运动目标，与专家目测屏幕扫描跟踪法相比，准确率达90％。同Amaud等人提出的面积重叠跟踪法相比，其准确率提高了一个多数量级。

This paper firstly introduces basic methods for identifying and tracking MCS ,and points out their advantages and disadventages. It then improves the algorithm of area overlapped method. Secondly, from the shape characters and essence attributes(e.g.textural attributes)of flow shaped group moving targets, the paper brings forward several kinds of method for constructing matching template. By considering the whole description, geometric shape and immovability moments characteristic of flow shaped group moving targets and its essence attributer, the templates matching is performed by using the principle of choosing minimum Eucliddistance, Meanwhile, the method of identifying and tracking targets is realized by using the arithmetic of matching across and updating elements of the matching templates in turn. The experiments show that the methods and skills can make a computer to automatically identify and track these multi flow shaped moving targets such as MCS . It is full of availability to identify and track th MCS by using them.