针对类圆形堆积物图像的前景和背景在色彩或灰度上相近，难以用传统阈值分割等算法进行有效分割的问题，提出一种多方法融

合的类圆形堆积物图像区域分割策略。对图像进行滤波等一系列预分割处理后进行投影得到目标图像的外接矩形区域，以排除噪声的干扰，

在区域内采用改进的灰度共生矩阵方法进行粗分割，以解决窗口大小与分割精度的问题，采用二维OTSU 阈值分割方法对粗分割结果进行

量化。实验结果表明，该方法得到的区域分割结果边缘清晰、准确度高。-Round deposits for the prospect of images and background color or grayscale at up close, it is difficult using traditional threshold segmentation algorithm, such as the effective partition of the problem, a method of integration of many types of deposits circular image region segmentation strategy. Image filtering deal with a series of pre-partition after the target image projection external rectangular area in order to rule out the possibility of noise interference, in the region to adopt improved methods of co-occurrence matrix gray coarse segmentation to solve the window size and segmentation accuracy problem, using two-dimensional OTSU threshold segmentation method to quantify the coarse segmentation results. Experimental results show that the method of region segmentation results of the edge of a clear, high accuracy.