基于智能视觉的综采面风险区域识别仿真  

Simulation of risk area recognition in fully mechanized mining face based on intelligent vision

作　　者：马力 王欢 贺建荣 项占锋 MA Li;WANG Huan;HE Jianrong;XIANG Zhanfeng(Shendong Intelligent Technology Service,Shaanxi 719315,China;Shaanxi Chuangjie Henglian Technology Co.,Ltd.,Shaanxi 710065,China)

机构地区：[1]国能神东煤炭智能技术中心,陕西719315 [2]陕西创杰恒联科技有限公司,陕西710065

出　　处：《自动化与仪器仪表》2025年第3期255-258,263,共5页Automation & Instrumentation

摘　　要：综采面环境复杂多变,对传统的安全监测手段提出了严峻的挑战。在此背景下,研究基于智能视觉的综采面风险区域识别仿真方法。获取综采面视觉图像获取并实施去噪和照度调节两种清晰化处理。计算视觉图像中的两种风险因子,即颜色因子、纹理因子。基于智能视觉中的AdaBoost算法构建分类器,以颜色因子、纹理因子为输入,利用该分类器实现综采面风险区域识别。结果表明:所研究的识别方法的ROC曲线下方AUC值相对更大,说明该识别方法具有更高的分类准确性。The complex and ever-changing environment of fully mechanized mining faces poses a serious challenge to traditional safety monitoring methods.In this context,a simulation method for identifying risk areas in fully mechanized mining faces based on intelligent vision is studied.Obtain visual images of the fully mechanized mining face and implement two types of clarity processing:denoising and illumination adjustment.Calculate two risk factors in visual images,namely color factor and texture factor.Build a classifier based on the AdaBoost algorithm in intelligent vision,using color and texture factors as inputs,and use this classifier to identify risk areas in fully mechanized mining faces.The results indicate that the AUC value below the ROC curve of the studied recognition method is relatively larger,indicating that the recognition method has higher classification accuracy.

关 键 词：智能视觉 综采面 风险区域 风险因子 识别方法 

分 类 号：TP55.2[自动化与计算机技术]