煤炭传送皮带表面裂纹检测技术的研究

发布时间：2018-04-02 19:30

本文选题：皮带裂纹检测　切入点：数字图像处理　出处：《辽宁科技大学》2015年硕士论文

【摘要】：胶带式输送机是许多大型煤矿广泛使用的输煤设备,其传送皮带是最重要的煤炭承载体,传送皮带一旦发生撕裂不仅会给安全生产带来极大威胁,而且会给生产单位带来巨大的经济损失。因此,提前预防传送皮带撕裂,检测传送皮带上出现的细小裂纹,具有重要的理论和实际工程意义。本文在分析了煤炭传送皮带的实际应用环境的基础上,根据物理学中漫反射原理,构造了传送皮带裂纹检测的整体结构,并根据工程需要选择了一款合适的CCD摄像机,并对图像进行了标定。针对传送皮带裂纹图像对比度低、存在噪声干扰等问题,先对采集到的裂纹图像进行图像增强处理。在此基础上,提出一种基于改进猴王遗传算法的Otsu图像分割算法,与基于遗传算法的Otsu分割方法相比,不仅缩短了算法的运行时间,并且很好的把裂纹从背景中分割出来。针对图像分割后裂纹图像可能存在孤立点和不连通的情况,利用数学形态学原理,构造菱形结构元素,该方法有效的消除掉了孤立点,并且使裂纹不连通的区域连通起来。裂纹图像本身可能存在交叠的情况,构造横向和纵向结构元素,很好的把交叠裂纹分割成横向和纵向裂纹,方便了特征提取时对几何参数的计算。对裂纹进行特征提取和识别时,先用骨架提取方法对裂纹图像进行细化,然后计算裂纹的长度和宽度,利用逐像素方法提取求出裂纹的面积和周长,最后根据圆形度准则判断是否为裂纹。
[Abstract]:Belt conveyor is widely used in many large-scale coal mine coal handling equipment, the conveyor belt is the most important coal bearing body, transmission belt once the tear will not only bring great threat to the safety in production, and will bring huge economic losses to the production units. Therefore, the prevention of transmission belt tearing in advance, small crack detection conveyor that has important theoretical and practical significance. Based on the analysis of the coal conveyor belt's practical application environment, based on the reflection of physics principle, structure of the overall structure of the transmission belt crack detection, and according to the engineering needs to choose a suitable CCD camera, and the image is calibrated. The conveyor belt crack image with low contrast, noise and other issues, to crack the collected images to enhance the image. On this basis, put forward An improved Otsu algorithm for image segmentation based on genetic algorithm, the monkey king, compared with the segmentation method based on genetic algorithm Otsu, not only shorten the running time of the algorithm, and put a good crack separated from the background. According to the image segmentation after the crack image may exist outliers and disconnected, using the principle of mathematical morphology. Structure of diamond structure elements, the method can eliminate the outliers, and the crack connected region connected together. The crack image itself may overlap, structure of transverse and longitudinal structural elements, good to the overlapping crack is divided into horizontal and vertical crack, calculation of the geometric parameters of the convenient features extraction. Feature extraction and recognition of the crack, first use the skeleton extraction method to refine the crack image, and then calculate the crack length and width, by using pixel method The area and perimeter of the crack are obtained, and the crack is judged according to the circle degree criterion.

【学位授予单位】：辽宁科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TD528.1

【相似文献】