基于图像对准式V棱镜折射仪低对比度图像的识别算法研究  

Algorithm Research on Image Recognition of Low Contrast Single Line in V-Prism Refractometer in Using Image Alignment

作　　者：李延风[1] 安志勇[1] 王劲松[1] 呼丹[1] 

机构地区：[1]长春理工大学光电工程学院,吉林长春130022

出　　处：《激光与光电子学进展》2014年第6期71-76,共6页Laser & Optoelectronics Progress

基　　金：吉林省科技发展计划重大专项(20126014)

摘　　要：高精度光学玻璃折射率是保证光学设计和成像质量的重要条件,主要由V棱镜折射仪进行检测。在图像对准式V棱镜折射仪中,用于对准的平行光管所成狭缝中的单线图像质量,直接影响折射率测量中的对准精度,尤其当单线与背景对比度不高时,会大大影响仪器的测量精度。提出一种自适应灰度拉伸和垂直投影相结合的图像增强算法,该方法能快速提取低对比度的单线图像。通过对比实验证明了本算法的有效性,将测角精度提高到了±1",算法稳定测量重复性优于1×10-6,对提高测量光学玻璃折射率的精度有实际意义。High precision refractivity of optical glass is an important condition to ensure the image quality, mainly by the V-prism refractometer detection. In V-prism refractometer using image alignment, the image quality of single line in the collimator directly affects the alignment accuracy in measurement, especially as the contrast between single line and background is low. A combination of self-adaptive gray equilibrium and vertical projection is proposed. This method can extract the low contrast single line, prove the effectiveness of the algorithm through the comparative experiments, and improve the angle measurement accuracy to ±1＂, repeatability superior to 1×10-6, and algorithm stability is good. It has important practical significance to achieve high precision measurement of optical glass refractive index.

关 键 词：图像处理 垂直投影 机器视觉 低对比度 图像增强 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]