针对透明物体逆光图像对比度低、可视质量差、部分区域过度曝光、边缘信息模糊等问题,提出了一种改进MSRCR(Multi-Scale Retinex with Color Restoration)的透明物体逆光图像增强算法.首先,在MSRCR算法的基础上添加最小可觉差的倒数作为光照分量的调节因子,解决图像的色偏问题,得到准确清晰的透明物体边缘信息;然后,利用自适应对比度增强算法对原图像进行处理,得到亮度适中,对比度高的图像;最后,将两幅图像按亮度均值比例进行拉普拉斯金字塔融合,并进行线性拉伸.将该算法应用于安瓿瓶视觉尺寸测量,结果表明:改进MSRCR的透明物体逆光图像增强算法,能有效解决MSRCR算法的色偏问题,突显透明物体边缘细节信息,并保留亮度增强效果,将安瓿瓶的尺寸误差由 0.35 mm降低到 0.1mm,提高了透明物体尺寸测量精度.
Improved MSRCR Algorithm for Backlight Image Enhancement of Transparent Objects
Targeting at the problems of low contrast,poor visual quality,overexposure of some areas and blurred edge information of transparent object backlight image,an improved MSRCR(Multi-Scale Retinex with Color Restoration)algorithm for transparent object backlight image enhancement is proposed.First of all,in order to solve the problem of color deviation of image,obtain accurate and clear edge information of transparent objects,reciprocal of the just noticeable difference is used as an adjustment factor for the illumination component of the MSRCR algorithm.Then,the adaptive contrast enhancement algorithm is used to process original image,and the image with moderate brightness and high contrast is obtained.Finally,the two images are fused with Laplacian pyramid fusion method accord-ing to the average luminance,and then stretched linearly.The algorithm is applied to vision measurement of ampoule dimensions.The results show that the improved MSRCR's transparent object backlight image enhancement algorithm can effectively solve the color devia-tion problem of MSRCR algorithm,highlight the edge details of transparent objects,retain the brightness enhancement effect,reduce the error of ampoule dimensions from 0.35 mm to 0.1 mm,and improve the accuracy of transparent object size measurement.
国家科技期刊平台
NSTL
万方数据
