In case of complex textures,existing static shadow detection and removal algorithms are prone to false detection of the pixels.To solve this problem,a static shadow detection and removal algorithm based on support vec...In case of complex textures,existing static shadow detection and removal algorithms are prone to false detection of the pixels.To solve this problem,a static shadow detection and removal algorithm based on support vector machine(SVM)and region sub-block matching is proposed.Firstly,the original image is segmented into several superpixels,and these superpixels are clustered using mean-shift clustering algorithm in the superpixel sets.Secondly,these features such as color,texture,brightness,intensity and similarity of each area are extracted.These features are used as input of SVM to obtain shadow binary images through training in non-operational state.Thirdly,soft matting is used to smooth the boundary of shadow binary graph.Finally,after finding the best matching sub-block for shadow sub-block in the illumination region based on regional covariance feature and spatial distance,the shadow weighted average factor is introduced to partially correct the sub-block,and the light recovery operator is used to partially light the sub-block.The experimental results show the number of false detection of the pixels is reduced.In addition,it can remove shadows effectively for the image with rich textures and uneven shadows and make a natural transition at the boundary between shadow and light.展开更多
The accuracy of extracting projected pig area is critical to the accuracy of the weight measurement of pigs by machine vision.The capability of both the conventional and the edge detection methods for extracting pig a...The accuracy of extracting projected pig area is critical to the accuracy of the weight measurement of pigs by machine vision.The capability of both the conventional and the edge detection methods for extracting pig area was examined using the images of 47 pigs of different weights.Relationship between the threshold value and the extracted area was numerically analyzed for both methods.It was found that the accuracy of the conventional method depended heavily on the threshold value,while choice of threshold value in the edge detection approach had no influence on the extracted area over a wide range.In normal lighting conditions,both methods yielded comparable values of predicted weight;however,under variable light intensities,the edge detection method was superior to the conventional method,because the former was proven to be independent of light intensities.This makes edge detection an ideal method for area extraction during the walk-through weighing process where pigs are allowed to move around.展开更多
针对复杂场景中背景复杂、目标周围噪声多及目标只占图像中较小部分而难于检测的问题,提出一种新的基于局部轮廓特征的检测目标方法.该方法首先利用改进的全局概率边界算法(Globalized probability of boundary,gPb)算法提取图像的轮廓...针对复杂场景中背景复杂、目标周围噪声多及目标只占图像中较小部分而难于检测的问题,提出一种新的基于局部轮廓特征的检测目标方法.该方法首先利用改进的全局概率边界算法(Globalized probability of boundary,gPb)算法提取图像的轮廓,然后应用最大类间方差法(Otsu)进行自动阈值处理得到图像的显著性轮廓;再提取显著性轮廓的k邻近大致直线轮廓段(k connected roughly straight contour segments,kAS),并以kAS作为局部特征,用于复杂场景中的目标检测.该算法结合gPb算法和Otsu提取轮廓的显著性轮廓,去除了目标附近的大量噪声边界,有效地提高了检测效率.同时,在检测阶段,测试集与训练集中提取的不相关特征数目也得到较大减少,从而提高了检测的精度.多组实验结果均表明本文方法的有效性.展开更多
提出多分辨奇异值分解(Multi-resolution singular value decomposition,MRSVD)的概念,基于矩阵二分递推构造原理,利用奇异值分解(Singular value decomposition,SVD)获得具有不同分辨率的近似和细节信号,以多分辨率来展现信号不同层次...提出多分辨奇异值分解(Multi-resolution singular value decomposition,MRSVD)的概念,基于矩阵二分递推构造原理,利用奇异值分解(Singular value decomposition,SVD)获得具有不同分辨率的近似和细节信号,以多分辨率来展现信号不同层次的概貌和细部特征。给出MRSVD的分解和重构算法,并从理论上证明这种分解方式的多分辨分析特性。研究结果表明,MRSVD可以精确地检测出信号中的奇异点位置,克服小波检测时的奇异点偏移缺陷,并具有优良的消噪能力,可实现零相移消噪,此外还具有微弱故障特征提取能力,在对一个轴承振动信号的处理中,提取到其中隐藏的周期性冲击特征,实现对轴承损伤的准确诊断。相应地与小波变换结果进行比较,证明MRSVD在信号处理和故障诊断领域是一种很有应用前景的方法。展开更多
基金University and College Scientific Research Fund of Gansu Province(No.2017A-026)Foundation of A hundred Youth Talents Training Program of Lanzhou Jiaotong University。
文摘In case of complex textures,existing static shadow detection and removal algorithms are prone to false detection of the pixels.To solve this problem,a static shadow detection and removal algorithm based on support vector machine(SVM)and region sub-block matching is proposed.Firstly,the original image is segmented into several superpixels,and these superpixels are clustered using mean-shift clustering algorithm in the superpixel sets.Secondly,these features such as color,texture,brightness,intensity and similarity of each area are extracted.These features are used as input of SVM to obtain shadow binary images through training in non-operational state.Thirdly,soft matting is used to smooth the boundary of shadow binary graph.Finally,after finding the best matching sub-block for shadow sub-block in the illumination region based on regional covariance feature and spatial distance,the shadow weighted average factor is introduced to partially correct the sub-block,and the light recovery operator is used to partially light the sub-block.The experimental results show the number of false detection of the pixels is reduced.In addition,it can remove shadows effectively for the image with rich textures and uneven shadows and make a natural transition at the boundary between shadow and light.
基金The project was supported in part by the National Research Initiative of the USDA Cooperative State Research,Education and Extension Service,grant number 2003-35503-13990.
文摘The accuracy of extracting projected pig area is critical to the accuracy of the weight measurement of pigs by machine vision.The capability of both the conventional and the edge detection methods for extracting pig area was examined using the images of 47 pigs of different weights.Relationship between the threshold value and the extracted area was numerically analyzed for both methods.It was found that the accuracy of the conventional method depended heavily on the threshold value,while choice of threshold value in the edge detection approach had no influence on the extracted area over a wide range.In normal lighting conditions,both methods yielded comparable values of predicted weight;however,under variable light intensities,the edge detection method was superior to the conventional method,because the former was proven to be independent of light intensities.This makes edge detection an ideal method for area extraction during the walk-through weighing process where pigs are allowed to move around.
文摘针对复杂场景中背景复杂、目标周围噪声多及目标只占图像中较小部分而难于检测的问题,提出一种新的基于局部轮廓特征的检测目标方法.该方法首先利用改进的全局概率边界算法(Globalized probability of boundary,gPb)算法提取图像的轮廓,然后应用最大类间方差法(Otsu)进行自动阈值处理得到图像的显著性轮廓;再提取显著性轮廓的k邻近大致直线轮廓段(k connected roughly straight contour segments,kAS),并以kAS作为局部特征,用于复杂场景中的目标检测.该算法结合gPb算法和Otsu提取轮廓的显著性轮廓,去除了目标附近的大量噪声边界,有效地提高了检测效率.同时,在检测阶段,测试集与训练集中提取的不相关特征数目也得到较大减少,从而提高了检测的精度.多组实验结果均表明本文方法的有效性.
文摘提出多分辨奇异值分解(Multi-resolution singular value decomposition,MRSVD)的概念,基于矩阵二分递推构造原理,利用奇异值分解(Singular value decomposition,SVD)获得具有不同分辨率的近似和细节信号,以多分辨率来展现信号不同层次的概貌和细部特征。给出MRSVD的分解和重构算法,并从理论上证明这种分解方式的多分辨分析特性。研究结果表明,MRSVD可以精确地检测出信号中的奇异点位置,克服小波检测时的奇异点偏移缺陷,并具有优良的消噪能力,可实现零相移消噪,此外还具有微弱故障特征提取能力,在对一个轴承振动信号的处理中,提取到其中隐藏的周期性冲击特征,实现对轴承损伤的准确诊断。相应地与小波变换结果进行比较,证明MRSVD在信号处理和故障诊断领域是一种很有应用前景的方法。
