众所周知，形状的表示与描述是模式识别的中心内容．然而，大多数简单类型的二维形状描述算法，无论是基于边界特性还是基于区域特性的，都有一定的应用范围或者在性能上存在某些不足．针对这样的情况，基于区域边界上当前像素与前后两个相邻像素的坐标关系，提出了一种称之为区域直方图的用于区域形状表示与描述的通用方法．该区域直方图是按照一定的分类与计算规则将区域沿边界序列化而得到的水平间距形式，它在像素意义上表示了区域的边界长度与面积．利用Rosen与Gleason推荐的标准二维形状进行了形状描述能力的定量测试以及同类型算法的对比试验，结果表明，该方法在区域形状描述方面显得很有效，能满足复杂区域形状的表示与描述要求，并且表现出较好的鲁棒性．它为区域形状的表示与描述提供了一种新的有效手段，从而在模式识别等机器视觉方面表现出一定的应用价值。

As we all know, Shape description is a central problem in pattern recognition. However, most simple algorithms for generating 2D shape descriptors, whether they are based on boundary features or regional features, can only be used in certain applications or have deficiencies in some performances. To address such problem, based on the coordinate relationship between the current boundary pixel and its two lockstep boundary pixels, a general-purpose method ( so called region histogram ) for region shape representing and describing is presented in this paper. Region histogram is the sequenced horizontal intervals along the region boundary in accordance with some classification and calculation criterion , which actually represent the region's area and its boundary length in terms of pixels. In order to quantify the recognition ability of the region histogram method and compare it with the analogous ones, a series of tests were carried out using a set of standard shapes, which have been proposed by Rosen and Gleason, comprising of a disc, a square, a rectangle and an equilateral. The results show that region histogram method is robust and effective to region shape representing and describing; even it does better with complex regions which have much more global and local contour information. Therefore, this method presents a new effective way for shape representing and describing, and can be used in pattern recognition such as robot vision applications.