《電子技術(shù)應(yīng)用》
您所在的位置:首頁(yè) > EDA與制造 > 設(shè)計(jì)應(yīng)用 > 非均勻光照下銅板表面缺陷圖像增強(qiáng)
非均勻光照下銅板表面缺陷圖像增強(qiáng)
電子技術(shù)應(yīng)用
楊凱鈞,陳逃
昆明理工大學(xué) 信息工程與自動(dòng)化學(xué)院
摘要: 針對(duì)銅板表面缺陷圖像容易受到非均勻光照影響,,出現(xiàn)反光和亮度失真,,導(dǎo)致圖像難以運(yùn)用到檢測(cè)中的問(wèn)題,提出了一種非均勻光照?qǐng)鼍跋裸~板表面缺陷圖像的增強(qiáng)方法,。首先提取圖像中的光照分量,然后將光照分量進(jìn)行分塊,根據(jù)不同塊的亮度進(jìn)行優(yōu)化,,并在分塊的基礎(chǔ)上進(jìn)行自適應(yīng)伽馬變換,調(diào)整圖像中的整體亮度,。然后,,使用Top-Hat變換加強(qiáng)圖像中的缺陷區(qū)域,最后將Top-Hat變換前后的圖像進(jìn)行融合,,得到最終的圖像,。實(shí)驗(yàn)結(jié)果表明,在擦傷,、劃痕,、孔洞這三類缺陷中,,增強(qiáng)后的圖像信息熵分別提升了10.51%、5.29%,、2.89%,,與其他圖像增強(qiáng)算法相比,所提算法能夠有效抑制銅板表面缺陷的反光,,提升圖像的質(zhì)量并增強(qiáng)圖像中的缺陷區(qū)域,。
中圖分類號(hào):TP751 文獻(xiàn)標(biāo)志碼:A DOI: 10.16157/j.issn.0258-7998.245044
中文引用格式: 楊凱鈞,陳逃. 非均勻光照下銅板表面缺陷圖像增強(qiáng)[J]. 電子技術(shù)應(yīng)用,,2024,,50(9):94-100.
英文引用格式: Yang Kaijun,Chen Tao. Enhancement of surface defect images on copper plates under non-uniform illumination[J]. Application of Electronic Technique,,2024,,50(9):94-100.
Enhancement of surface defect images on copper plates under non-uniform illumination
Yang Kaijun,Chen Tao
Faculty of Information Engineering and Automation,, Kunming University of Science and Technology
Abstract: Aiming at the problem that the image of defects on the surface of copper plate is easily affected by non-uniform illumination, which results in reflections and brightness distortion and makes it difficult to apply the image to detection, an enhancement method for the image of defects on the surface of copper plate under non-uniform illumination scenario is proposed. Firstly, the light components in the image are extracted, chunked and then optimized according to the brightness of different chunks. Based on the chunking, adaptive gamma correction is performed to adjust the overall brightness of the image. Then, Top-Hat transform is used to enhance the defective regions in the image, and finally the images before and after Top-Hat transform are fused to obtain the final image. The experimental results show that the information entropy of the enhanced image is improved by 10.51%, 5.29%, and 2.89% in three types of defects, namely scratches, scratches, and holes, respectively. Compared with other image enhancement algorithms, the proposed algorithm can effectively inhibit the reflection of the defects on the surface of the copper plate, improve the quality of the image and enhance the defective regions in the image.
Key words : image enhancement,;multiscale fusion;gamma correction,;non-uniform illumination

引言

在工業(yè)制造和表面缺陷檢測(cè)領(lǐng)域,,銅板作為重要材料,廣泛應(yīng)用于多個(gè)領(lǐng)域,。隨著檢測(cè)技術(shù)的發(fā)展,,利用機(jī)器視覺對(duì)銅板表面缺陷進(jìn)行檢測(cè)的應(yīng)用也在逐漸擴(kuò)大。然而,,在實(shí)際應(yīng)用中,,由于非均勻光照影響,銅板表面高光噪聲成為缺陷檢測(cè)的主要障礙之一[1],,并且銅板表面的圖像往往呈現(xiàn)出高反光,、亮度失真等問(wèn)題,從而顯著降低了圖像的質(zhì)量,。

目前,,針對(duì)光照不均勻的圖像增強(qiáng)主要分為基于空間域[2-4]和基于頻率域[5-7]兩大方面?;诳臻g域的圖像增強(qiáng)方法包括直方圖均衡化方法和Retinex方法等,。基于頻率域的方法主要有:低通濾波,、高通濾波,、同態(tài)濾波等。直方圖均衡化[8]通過(guò)重新分布圖像的灰度級(jí)來(lái)增強(qiáng)對(duì)比度,,它對(duì)整體圖像進(jìn)行處理,。Retinex方法[9]受啟發(fā)于人眼對(duì)光照的適應(yīng)性,,SSR和MSR是常見的Retinex變體?;陬l率域的圖像增強(qiáng)算法將圖像視為包含高頻和低頻復(fù)合信號(hào)[10],。傅里葉變換[11]是一種常見的方法,但此類算法通常在自適應(yīng)性方面較為有限,。孫闊原等[12]采用最大類間方差法進(jìn)行值分割,,結(jié)合形態(tài)學(xué)濾波及邊緣幾何特征對(duì)缺陷進(jìn)行定位檢測(cè)。王偉江等人[13]針對(duì)光照問(wèn)題提出基于形態(tài)學(xué)熵圖像的光照歸一化算法,,能有效增強(qiáng)光照不均圖像,。王凡等人[14]利用多結(jié)構(gòu)形態(tài)學(xué)增強(qiáng)算法可以有效提取光照不均圖像中形狀較小的缺陷。劉志成等[15]提出了一種將Retinex理論與伽馬變換相結(jié)合的自適應(yīng)調(diào)整算法,,可以保留更詳細(xì)的信息,但仍會(huì)存在亮度飽和效應(yīng),。湯子麟等人[16]構(gòu)造了一種自適應(yīng)伽馬矯正函數(shù),,可以兼顧圖像的全局特性和局部細(xì)節(jié)信息,但對(duì)強(qiáng)光部分處理效果較差,。深度學(xué)習(xí)方法,,如RetinexNet[17]和AM-RetinexNet [18]將Retinex與神經(jīng)網(wǎng)絡(luò)相結(jié)合,使用大量數(shù)據(jù)集訓(xùn)練校正模型,,在圖像降噪和色彩視覺修正等方面表現(xiàn)較好,,但對(duì)圖像中的反光區(qū)域優(yōu)化效果較差。

上述的研究對(duì)改善圖像質(zhì)量具有一定的效果,,但還是存在一些不足之處,,在優(yōu)化圖像的過(guò)程中,會(huì)影響圖像中的正常部分,,從而失去圖像中的部分細(xì)節(jié),,且對(duì)反光部分優(yōu)化較差。針對(duì)這些問(wèn)題,,本文提出了一種非均勻光照?qǐng)鼍跋裸~板表面缺陷圖像的增強(qiáng)方法,。


本文詳細(xì)內(nèi)容請(qǐng)下載:

http://forexkbc.com/resource/share/2000006149


作者信息:

楊凱鈞,陳逃

(昆明理工大學(xué) 信息工程與自動(dòng)化學(xué)院,,云南 昆明 650504)


Magazine.Subscription.jpg

此內(nèi)容為AET網(wǎng)站原創(chuàng),,未經(jīng)授權(quán)禁止轉(zhuǎn)載。