一种图像分割的快速目标描述方法

An Approach of Fast Object Description for Image Segmentation

提出了一种二值图像分割的快速目标描述算法。该方法以像素标记法为基础,利用相邻扫描行中二值像素点的邻接关系,在一次扫描的过程中完成目标标记的提取以及等价目标标记的合并,最终完成目标的标记和特征量的提取工作,具有速度快,方法简单,处理复杂图案能力强的特点。

A new approach of object description after image segmentation to obtain objects’ feature is introduced. It reduce the numbers of image scanning to 1 time and combine equivalent label in image scanning. All objects’ feature can be obtained in the time of scanning finishing. The corresponding label information is update when a new pixel is scanned. All equivalent labels are combined as soon as possible when label is updating. No more time is needed after pixel scan finished. The flow of this approach is listed step by step with particular description. A complex example is analyzed with picture. Correctness and performance of this approach is proved with all possible instances are illuminated.