工作面电磁波ct图像的自动识别与解释系统安全技术及工程专业论文.docxVIP

一f61605硕士学位论文 一f61605 硕士学位论文 摘要 摘 要 准确查清工作面地质异常关系着矿井安全生产。目前在工作面电磁波CT探 测中，对CT图像的解释主要是人工解释为主，影响的因素较多，往往造成解 释出现偏差。本论文主要研究数字图像处理技术和人工神经网络在工作面电磁 波CT图像自动识别和解释中的应用。主要探讨了图像平滑、中值滤波、形态 学变换等算法应用于工作面电磁波CT图像处理，并分析了处理的效果；通过 图像分割技术去除了CT图像中地质条件正常的区域，利用局部邻域算法进行 连通区域标记，从而进一步提取CT图像的面积、周长、圆度等特征：通过构 造BP神经网络分类器进行模式识别，实现了工作面电磁波CT图像的自动识别 与解释，并结合应用实例进行系统运行试验。 关键字：工作面，电磁波CT，图像处理，模式识别，特征提取，地质构造 Abstract To detect geological anomaly accurately concerns safety coal mining At present the electromagnetic wave computerized tomography(CT) technique is used to detect geological anomaly of working face，but the interpretation for the CT image is mainly of manual interpretation，which is often affected by many factors，thus appears deviation．This paper chiefly studies the applications of digital image processing technology and artificial neural network in the automatic identification and interpretation of the electromagnetic wave CT image of working face．It mainly discusses the algorithms，such as image smoothing，median filter and morphology transformation，which are applied to electromagnetic wave CT image of working face processing，and analyzes the processing 硕士学位论文 硕士学位论文 摘要 effectiveness．The region with normal geological conditions in the CT image is removed by means of image segmentation，and the connected region is marked by partial neighborhood algorithm，then the features of CT image，such as area，perimeter and roundness are extracted．By creating categorizer of back-propagation(BP)neural network to make pattern recognition，the automatic recognition and interpretation of electromagnetic wave CT of working face are realized，and an example iS tested on the system． Keywords：working face，electromagnetic wave computerized tomography,image processing，pattem recognition，feature extraction， geological structure 堡主兰丝堡塞翌』羔堂第1章绪论 堡主兰丝堡塞翌』羔堂 第1章绪论 1．1