降落岩石骨料图像阈值分割算法

岩石破碎质量可通过对皮带末端降落过程中岩块摄像图像的分析来实现连续监测。因此，摄像图像的处理是该系统的关键问题。为此，研究了一种降落岩石骨料灰度图像的局部阈值二值化算法，该算法重复分割岩石骨料图像直到图像中所有目标分割完毕。目标分割前，在每一次迭代分割过程中计算目标的尺寸、形状以及其灰度变化范围。本算法已应用于在线系统，相关测试结果表明在一定的条件下分割效果较好。

Thresholding Algorithm for Aggregate Images from a Gravity Flow

Rock crushing quality can be continuous monitored by means of photographing the aggregates falling from the end of conveyer belt. Here, the image treatment is a key point. For this, a local thresholding algorithm for binarizing the gray scale images of falling aggreates is developed. The algorithm repeatedly thresholds an aggregate image until no object can be thresholded any further. Before thresholding an object, in each iteration of the thresholding process, the algorithm calculates the size, shape and gray level range of the object. The algorithm has been implemented in an on line system and the related test results show it works well under certain conditions.