摘要
放顶煤煤矸识别影响着整个综放工作面的开采效率,可以解决欠放和过放问题,使工人远离放煤工作面。提出基于神经网络图像识别技术的放顶煤煤矸自动识别方法。结合多阈值Qtsu分割算法和边缘检测方法,将图像分割为煤矸层、烟煤层和背景层,在煤矸层中利用灰度共生矩阵提取图像特征,在神经网络中加入模糊补偿原理,将提取的特征输入优化后的神经网络中,完成放顶煤煤矸自动识别。实验结果表明,所提方法的分割精度高、识别精度高、复杂度低。
The identification of coal gangue in top coal caving affects the mining efficiency of the whole fully mechanized top coal caving face.It can solve the problems of under caving and over caving,and make the workers away from the coal caving face.An auto-matic recognition method of coal gangue in top coal caving based on neural network image recognition technology is proposed.Combined with the multi threshold qtsu segmentation algorithm and the edge detection method,the image is divided into coal gangue layer,bituminous coal layer and background layer.In the coal gangue layer,the gray level co-occurrence matrix is used to extract the image features,the fuzzy compensation principle is added to the neural network,and the extracted features are input in-to the optimized neural network to complete the automatic recognition of the top coal gangue.Experimental results show that the proposed method has high segmentation accuracy,high recognition accuracy and low complexity.
作者
范忠明
FAN Zhong-ming(China Energy Wuhai Energy Co.,Ltd.,Inner Mongolia Autonomous Region,Wuhai 016000 China)
出处
《自动化技术与应用》
2024年第10期39-42,共4页
Techniques of Automation and Applications
基金
乌海能源项目(WHNY-KX-21-05)。
关键词
神经网络
图像分割
放顶煤煤矸识别
边缘检测
模糊补偿原理
特征提取
neural network
image segmentation
top coal gangue recognition
edge detection
fuzzy compensation principle
feature extraction
