摘要
煤岩图像识别是实现采掘工作面无人化的基础。研究了字典学习法、小波变换法、灰度共生矩阵法等主流算法在煤岩图像识别应用中的适用范围和存在的问题。提出了基于多参数融合的煤岩识别方法:提取温度、声音、振动、粉尘浓度、图像等特征参量,结合各自的优点,采用深度学习等先进技术,能够有效提高煤岩图像的鲁棒性及识别率。
Coal-rock image recognition is the basis for unmanned mining face.In this paper,the application scope and problems of dictionary learning method,wavelet transform method and gray symbiosis matrix method in coal rock image recognition are studied.A coal-rock identification method based on multi-parameter fusion is proposed:extracting characteristic parameters such as temperature,sound,vibration,dust concentration,image,etc.,combined with their respective advantages,using advanced techniques such as deep learning to effectively improve the robustness and recognition rate of coal-rock images.
作者
王莹
Wang Ying(School of Mechatronics Electronic and Information Engineering,China University of Mining and Technology(Beijing),Beijing 100083,China)
出处
《能源与环保》
2019年第4期139-142,共4页
CHINA ENERGY AND ENVIRONMENTAL PROTECTION
关键词
煤岩图像识别
无人化
多参数
深度学习
鲁棒性
coal-rock image recognition
unmanned
multi-parameter
deep learning
robustness
