基于光照校正和图像融合的零件表面图像增强

doi:10.19678/j.issn.1000-3428.0067684

摘要/Abstract

摘要： 在低照度条件下采集的机械加工零件表面图像质量较低,严重影响后续与粗糙度相关的特征提取效果。为此,提出一种基于光照校正和图像融合的低照度图像增强算法。首先对引导滤波算法进行改进,使其能根据图像的纹理情况自适应地调整滤波参数,在平滑纹理的同时保持整体结构,得到更优质的照度图;然后对原始图像进行反相增强,抑制图像中的高亮条纹与光斑,通过图像融合来综合原图和正、反相增强图像的优质像素;接着利用限制对比度的自适应直方图均衡化(CLAHE)算法进一步提高融合图像的对比度;最后将增强图像用于2种不同的零件表面粗糙度检测模型。在钛、钢、镁3种材质的零件表面低照度图像数据集上进行实验,结果表明,该算法有效增强了低照度零件表面图像的质量,增强图像的标准差、平均梯度和信息熵均高于对比算法。与图像增强前相比,基于灰度共生矩阵特征提取的支持向量回归(GLCM-SVR)和基于卷积回归神经网络(RCNN)的粗糙度检测模型的均方根误差分别降低了0.140和0.202 μm,平均绝对误差分别降低了0.116和0.146 μm,表明图像增强能有效提升低照度条件下基于视觉的粗糙度检测方法的精度。

关键词: 表面粗糙度, 图像增强, Retinex理论, 光照校正, 图像融合

Abstract: The quality of surface images of machined parts collected under low illumination conditions is relatively poor, which significantly affects the subsequent extraction of roughness-related features. To this end, a low-light image enhancement algorithm based on illumination correction and image fusion is proposed. First, the guided filtering algorithm is improved to adaptively adjust the filtering parameters based on the texture of the image. This technique maintains the overall structure while smoothing the texture and obtains a higher quality illumination map. Then, inverse enhancement is performed on the original image to suppress the bright stripes and spots and integrate the high-quality pixels of the original image and the positive and negative enhanced images through image fusion. Subsequently, the Contrast Limited Adaptive Histogram Equalization (CLAHE) algorithm with limited contrast is used to enhance the contrast of the fused image. Finally, two different surface roughness detection models of parts are applied to the enhanced images. Experiments are conducted on a low-light image dataset of titanium, steel, and magnesium parts. The results show that the algorithm effectively enhances the quality of low-light part surface images, and the standard deviation, average gradient, and information entropy of the enhanced images are higher than those obtained with the existing algorithm. Compared with the results before image enhancement, the root mean square errors of the roughness detection models based on Gray Level Co-occurrence Matrix with Support Vector Regression (GLCM-SVR) and Regression Convolutional Neural Network (RCNN) are reduced by 0.140 and 0.202 μm, and the average absolute errors decrease by 0.116 and 0.146 μm, respectively. This indicates that image enhancement can effectively improve the accuracy of vision-based roughness detection methods under low illumination conditions.

Key words: surface roughness, image enhancement, Retinex theory, illumination correction, image fusion

中图分类号:

TP391.41

董良振, 田建艳, 杨胜强, 陈海滨. 基于光照校正和图像融合的零件表面图像增强[J]. 计算机工程, 2024, 50(6): 245-254.

DONG Liangzhen, TIAN Jianyan, YANG Shengqiang, CHEN Haibin. Image Enhancement of Parts Surface Based on Illumination Correction and Image Fusion[J]. Computer Engineering, 2024, 50(6): 245-254.

https://www.ecice06.com/CN/Y2024/V50/I6/245

参考文献

[1] 张存鹰,赵波.超声振动辅助加工表面微结构及其特性研究进展[J].表面技术, 2019, 48(5):259-274. ZHANG C Y, ZHAO B. Research progress of properties of surface micro-structure in ultrasonic vibration assisted machining[J]. Surface Technology, 2019, 48(5):259-274.(in Chinese)
[2] 陈苏婷,胡海锋,张闯.基于激光散斑成像的零件表面粗糙度建模[J].物理学报, 2015, 64(23):113-121. CHEN S T, HU H F, ZHANG C. Surface roughness modeling based on laser speckle imaging[J]. Acta Physica Sinica, 2015, 64(23):113-121.(in Chinese)
[3] PATEL D R, KIRAN M B, VAKHARIA V. Modeling and prediction of surface roughness using multiple regressions:a noncontact approach[J]. Engineering Reports, 2020, 2(2):1-15.
[4] 杨洁,李乐.基于机器视觉的表面粗糙度测量与三维评定[J].光学技术, 2016, 42(6):491-495. YANG J, LI L. Surface roughness measurement and three-dimensional assessment based on machine vision[J]. Optical Technique, 2016, 42(6):491-495.(in Chinese)
[5] 魏万珍.基于图像技术的零件表面粗糙度检测方法及系统研究[D].太原:太原理工大学, 2022. WEI W Z. Research on detection method and system of surface roughness of parts based on image technology[D].Taiyuan:Taiyuan University of Technology, 2022.(in Chinese)
[6] ZHAO L, LU S P, CHEN T, et al. Deep symmetric network for underexposed image enhancement with recurrent attentional learning[C]//Proceedings of IEEE/CVF International Conference on Computer Vision. Washington D.C.,USA:IEEE Press,2021:12055-12064.
[7] LI J J, FENG X M, HUA Z. Low-light image enhancement via progressive-recursive network[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2021, 31(11):4227-4240.
[8] 文海琼,李建成.基于直方图均衡化的自适应阈值图像增强算法[J].中国集成电路, 2022, 31(3):38-42, 71. WEN H Q, LI J C. An adaptive threshold image enhancement algorithm based on histogram equalization[J]. China Integrated Circuit, 2022, 31(3):38-42, 71.(in Chinese)
[9] 艾青松,张皓喆,严俊伟.基于自适应伽马校正的异常驾驶行为检测方法[J].计算机工程, 2023, 49(9):279-286. AI Q S, ZHANG H Z, YAN J W. Abnormal driving behavior detection method based on adaptive gamma correction[J]. Computer Engineering, 2023, 49(9):279-286.(in Chinese)
[10] BHANDARI A K, SRINIVAS K, MAURYA S. Gamma corrected reflectance for low contrast image enhancement using guided filter[J]. Multimedia Tools and Applications, 2022, 81(4):6009-6030.
[11] 翟海祥,何嘉奇,王正家,等.改进Retinex与多图像融合算法用于低照度图像增强[J].红外技术, 2021, 43(10):987-993. ZHAI H X, HE J Q, WANG Z J, et al. Improved Retinex and multi-image fusion algorithm for low illumination image enhancement[J]. Infrared Technology, 2021, 43(10):987-993.(in Chinese)
[12] 丁元,邬开俊.基于RGB色彩平衡方法的沙尘降质图像增强[J].光学精密工程, 2023, 31(7):1053-1064. DING Y, WU K J. Sand-dust image enhancement using RGB color balance method[J]. Optics and Precision Engineering, 2023, 31(7):1053-1064.(in Chinese)
[13] JOBSON D J, RAHMAN Z, WOODELL G A. Properties and performance of a center/surround Retinex[J]. IEEE Transactions on Image Processing, 1997, 6(3):451-462.
[14] LIN H N, SHI Z W. Multi-scale Retinex improvement for nighttime image enhancement[J]. Optik, 2014, 125(24):7143-7148.
[15] DONG L L, ZHAO L J, WANG J. Image enhancement via texture protection Retinex[J]. IET Image Processing, 2022, 16(1):61-78.
[16] LECCA M. A Retinex inspired bilateral filter for enhancing images under difficult light conditions[C]//Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications. Washington D.C.,USA:IEEE Press,2021:76-86.
[17] MU Q, WANG X Y, WEI Y Y, et al. Low and non-uniform illumination color image enhancement using weighted guided image filtering[J]. Computational Visual Media, 2021, 7(4):529-546.
[18] 周友行,易倩,杨文佳,等.改进DenseNet模型在工件表面粗糙度视觉检测中的应用[J/OL].机械科学与技术:1-6[2023-04-07].https://doi.org/10.13433/j.cnki.1003-8728.20230010. ZHOU Y X, YI Q, YANG W J, et al. Application of improved DenseNet model in visual inspection of workpiece surface roughness[J]. Mechanical Science and Technology for Aerospace Engineering:1-6[2023-04-07].https://doi.org/10.13433/j.cnki.1003-8728.20230010.(in Chinese)
[19] 易怀安,陆玲莉,舒爱华,等.基于神经网络和仿真数据的粗糙度分类检测[J].组合机床与自动化加工技术, 2022(12):72-76, 80. YI H A, LU L L, SHU A H, et al. Roughness classification and detection based on neural network and simulation data[J]. Modular Machine Tool&Automatic Manufacturing Technique, 2022(12):72-76, 80.(in Chinese)
[20] LAND E H, MCCANN J J. Lightness and Retinex theory[J]. Journal of the Optical Society of America, 1971, 61(1):1-11.
[21] 田会娟,蔡敏鹏,关涛,等.基于YCbCr颜色空间的Retinex低照度图像增强方法研究[J].光子学报, 2020, 49(2):0210002. TIAN H J, CAI M P, GUAN T, et al. Low-light image enhancement method using Retinex method based on YCbCr color space[J]. Acta Photonica Sinica, 2020, 49(2):0210002.(in Chinese)
[22] 刘杰,张建勋,代煜.基于多引导滤波的图像增强算法[J].物理学报, 2018, 67(23):293-302. LIU J, ZHANG J X, DAI Y. Image enhancement based on multi-guided filtering[J]. Acta Physica Sinica, 2018, 67(23):293-302.(in Chinese)
[23] 贺钰茹.基于深度学习与MRI结构信息的动态PET滤波方法研究[D].广州:南方医科大学, 2021. HE Y R. Dynamic PET image filtering method based on deep learning and MRI structure information[D].Guangzhou:Southern Medical University, 2021.(in Chinese)
[24] XU L, YAN Q, XIA Y, et al. Structure extraction from texture via relative total variation[J]. ACM Transactions on Graphics, 2012, 31(6):139.
[25] ALI U, LEE I H, MAHMOOD M T. Guided image filtering in shape-from-focus:a comparative analysis[J]. Pattern Recognition, 2021, 111:107670.
[26] 张红英,朱恩弘,吴亚东.一种基于细节层分离的单曝光HDR图像生成算法[J].自动化学报, 2019, 45(11):2159-2170. ZHANG H Y, ZHU E H, WU Y D. High dynamic range image generating algorithm based on detail layer separation of a single exposure image[J]. Acta Automatica Sinica, 2019, 45(11):2159-2170.(in Chinese)
[27] LIU S G, ZHANG Y. Detail-preserving underexposed image enhancement via optimal weighted multi-exposure fusion[J]. IEEE Transactions on Consumer Electronics, 2019, 65(3):303-311.
[28] MERTENS T, KAUTZ J, VAN REETH F. Exposure fusion:a simple and practical alternative to high dynamic range photography[J]. Computer Graphics Forum, 2009, 28(1):161-171.
[29] 杨卫中,徐银丽,乔曦,等.基于对比度受限直方图均衡化的水下海参图像增强方法[J].农业工程学报, 2016, 32(6):197-203. YANG W Z, XU Y L, QIAO X, et al. Method for image intensification of underwater sea cucumber based on contrast-limited adaptive histogram equalization[J]. Transactions of the Chinese Society of Agricultural Engineering, 2016, 32(6):197-203.(in Chinese)

选择文件类型/文献管理软件名称

选择包含的内容