基于3DFR算法的爆破块度图像处理研究及其应用

Research and application on blasting fragmentation image processing based on 3DFR algorithm

考虑地下矿山拍照时的低对比度、噪声的干扰及灰度的不确定性等因素的特殊条件,提出适合地下矿山大块率测定的3DFR算法图像处理技术:首先,根据概率论确定三维直方图中目标与背景的区域;然后,通过模糊熵求极大值原理选取三维阈值,进而对图像进行处理;选取9组关于爆破块度的照片进行分割处理,并与OTSU法和传统方法进行分割结果对比评价。为了进一步考察3DFR算法图像处理技术的准确性,与现场实际测量结果进行对比和相关性分析。研究结果表明:3DFR算法图像处理结果与现场实际测量结果较吻合,相关性系数达到0.955,为地下矿山爆破大块率安全、准确、高效的测定提供了一条途径,并可用于地下矿山爆破参数优化。

Considering the problems of the low contrast, noise interference and gray uncertainty appears when taking pictures of the underground mine, the processing technology of 3DFR algorithm was put forward to suit the rock fragment rate in the underground mine. The procedures were as follows： Firstly, the target and setting in three-dimensional histogram were determined according to the probability theory. Then, according to the fuzzy entropy and the maximum principle, three-dimensional threshold was selected, and the images were disposed. Nice groups of blasting fragment photos were selected for analysis by OTSU method, traditional method and 3DFR method. Comparison and correlation analysis between the analysis results and the measured results were deduced to study the veracity of the processing technology of 3DFR algorithm. The results show that the correlation coefficient between the analysis results and the measured results is 0.955, which provides a way for safe, accurate and efficient determination of the blasting fragment rate of the underground mine and provides guidance for optimization of blasting parameters of the underground mine.