矿岩爆破图像二值化分割技术的研究与选择  

Study and Optimization of Rock Blasting Image Segmentation Technology based on Binaryzation

作　　者：史秀志[1] 黄丹[1] 盛惠娟[2] 周健[1] 赵建平[1] 

机构地区：[1]中南大学资源与安全工程学院,长沙410083 [2]中南大学地球科学与信息物理学院,长沙410083

出　　处：《爆破》2014年第1期47-50,113,共5页Blasting

基　　金：国家自然科学基金(51074178);中南大学中央高校基本科研业务费专项资金资助(2012zzts103)

摘　　要：爆破块度图像分析法是当前爆破块度统计的重要方法,图像分割技术是该分析方法的关键与难点。针对爆堆及块石图像的特殊性,研究了基于二值化的矿岩爆破图像分割技术,分析了阈值分割、区域生长分割、标记分水岭分割等当前主要的图像分割方法。通过研究图像上同一截面灰度值切片的变换与矿块实际边缘的吻合程度,发现数学形态学分割方法得到的二值图像边界清晰、断点少,相对于门限法更适用于爆破块度图像,其生成的二值化图像层次清晰,便于像素特征的提取。图像分割先进技术的应用,对于实现矿岩爆破图像块度分析的智能化提供了广阔的空间。Blasting fragmentation image analysis was an important method in the statistics on blasting rock frag- ment-size currently, while the image segmentation technique was the key issue. According to the specialty of the rock blasting image, the main binaryzation methods of image segmentation, such as threshold segmentation, region grow al- gorithm and marked watershed algorithm were discussed. By studying the extent to which the transform of gray value fitting the real edge of rock particle at the same cross section, the binary image generated by mathematical morpholo- gy algorithm showed a clear grain on which the pixel feature could be easily extracted. The experimental results showed that the mathematical morphology algorithm was more suitable than threshold method to segment blasting fragmentation image, which provided a broad space to realize the intelligent of rock blasting fragmentation analysis.

关 键 词：爆破块度 爆破图像 图像分割技术 二值化 数学形态学 

分 类 号：TD235.46[矿业工程—矿井建设]