中文摘要：

目的针对煤矿皮带堆煤传感器受复杂环境干扰导致检测效率较低的问题，提出一种新的基于熵能量（EE）的堆煤图像定位与识别方法。方法该方法首先在堆煤图像初始化区域内设定采样单元，统计采样单元前景和背景区域的信息熵，计算信息熵分布的极大值所对应的的灰度阶并统计其出现的概率；然后计算采样单元的熵能量和能量频率，归一化频率系数，并利用能量频率的二阶偏导确定堆煤边缘区域，将采样单元移动至堆煤边缘处；最后，根据采样单元的坐标位置计算堆煤高度和半径，进行实时预警。结果在煤矿井下皮带堆煤视频上进行算法测试，该算法的单帧图像平均处理速度为0．35s／帧，堆煤事故预警时间间隔最大为0．35s，平均识别准确率为97．1％，测试结果表明，该方法对堆煤模糊边缘、环境噪音具有抗干扰特点，可以有效克服强噪音、模糊边缘等因素对堆煤识别的影响。结论该方法对复杂环境中堆煤定位与识别具有很好的实时性和准确性，对提高煤矿安全生产效率具有重要意义。

英文摘要：

Objective A novel approach to coal pile location and recognition, which is based on entropy energy, is proposed in this paper to solve interference problems in a complex environment, which inhibit detection efficiency in coal mines. Method First, continuous information sampling units are scattered in the image, and the distribution area and gray levels of the larger probability density in the sampling unit are determined. All energy frequencies of the sampling unit are then cal- culated, the weakest of which is removed by filtration, and the edge of coal pile is calculated by second-order partial deriva- tives. The sampling unit is moved to the edge of the coal pile. Finally, the early warning based on the coal pile height and radius is ready for the upper limit, which is calculated according to the sampling unit. Result The average processing speed of the single image is O. 35 seconds per frame, the maximum of coal accident warning time interval is 0. 35 seconds, and the average recognition accuracy is 97. 1 percent, which is the method text results on the coal pile videos of coal mines. Experiments show that. This method can overcome the effects of noise and blurred image on the recognition of coal pile by taking advantage of the ability of entropy energy to identify a fuzzy edge and its resistance to environmental noise. Conclusion Experi- ments show that the proposed approach has good adaptability to coal pile location and recognition, and is of significant and practical value to improve the efficiency and safety of coal mine production.