光栅表面缺陷识别检测算法及应用研究

发布时间：2018-04-02 02:01

本文选题：光栅　切入点：图像处理　出处：《哈尔滨理工大学》2017年硕士论文

【摘要】：近年来,编码器被广泛的应用在各行业,市场的需求量大比重增加,而这也促进了编码器生产商的技术水平不断完善。编码器是使用圆光栅片做信号存储元件,测量精度高、运行状态平稳性好、可使用区间较广、占用空间小、质量轻、功率低和成本低等优点,其质量直接影响到使用性能。由于光栅片制造工艺的限制,其表面往往会产生星点、铬点、毛刺、断线等缺陷。目前,光栅表面缺陷的检测方式主要以人工检测为主,劳动强度大、主观性强、成本高,难以在精度、效率及自动化程度等方面满足检测要求。因此,本文应用图像处理技术,开发了光栅表面缺陷检测系统。首先,根据项目技术要求,对光栅夹具、二维平移台等机械结构进行了改进设计,并优化了控制系统参数,提高了系统的可靠性及控制精度,为稳定光栅图像的获取提供了保证。其次,针对光栅图像特点,进行了图像去噪、图像增强及运动模糊图像复原等图像预处理算法研究,并分析了常用的图像边缘重绘、边缘检测算法,提出了一种基于Canny算子的改进边缘重绘算法,保证了光栅表面轮廓特征的提取。再次,对预处理后的图像进行ROI截取,得到有效待检区域,并提取目标缺陷区域的几何特征,同时完成参数测量,实现了多种光栅表面缺陷的特征识别,并对缺陷区域进行标记。最后,设计了交互键盘、自动警示装置,并对软件系统进行了优化,提高了检测系统的可应用性。通过应用检测测试,发现检测系统的系统稳定性和检测效率等都有显著提高,实验表明,针对直径范围为33-38mm光栅,本文开发的光栅表面检测系统缺陷识别率在95%以上,误检率在5%以下,检测效率达29秒/个。
[Abstract]:In recent years, encoders have been widely used in various industries, the market demand has increased, and this has promoted the continuous improvement of the technical level of encoder manufacturers. It has the advantages of good stability, wide range, small space, light weight, low power and low cost. Its quality directly affects its performance. Due to the limitation of the fabrication process of the grating sheet, the surface of the grating will often produce stars. At present, the inspection methods of grating surface defects are mainly manual inspection, high labor intensity, strong subjectivity, high cost, difficult in precision, Efficiency and automation meet the requirements of detection. Therefore, this paper uses image processing technology to develop a grating surface defect detection system. First, according to the project technical requirements, the grating fixture, The mechanical structure of the two-dimensional translation platform is improved and the parameters of the control system are optimized. The reliability and control accuracy of the system are improved, and the stability of the grating image is guaranteed. Secondly, according to the characteristics of the grating image, The image pre-processing algorithms such as image denoising, image enhancement and motion blur image restoration are studied, and the commonly used edge redrawing and edge detection algorithms are analyzed. An improved edge redrawing algorithm based on Canny operator is proposed. Thirdly, the pre-processed image is intercepted by ROI to obtain the effective region to be checked, and the geometric feature of the target defect region is extracted, at the same time, the parameter measurement is completed. Finally, the interactive keyboard and automatic warning device are designed, and the software system is optimized. The system stability and detection efficiency of the detection system are significantly improved by applying the test results. The experimental results show that the diameter range is 33-38mm grating. The defect recognition rate of the grating surface detection system is over 95%, the error detection rate is less than 5%, and the detection efficiency is up to 29 seconds per second.
【学位授予单位】：哈尔滨理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TN25

【相似文献】