一种新的基于颜色恒常性的肤色分割算法

为了解决不同光照条件下的肤色分割问题,提出了一种新颖并且行之有效的解决方法。该方法先应用颜色恒常性算法对图像光照进行估计,根据光照估计的结果,在YCbCr颜色空间采用不同分布的高斯肤色模型,并利用动态阈值对图像进行分割。试验结果证明了算法的可行性和优越性。     

基于BP神经网络的颜色补偿模型

针对光源渐变等因素在机器视觉中产生的相关问题,提出了一种基于BP神经网络的图像颜色校正方法.该方法通过合适的训练集对BP神经网络进行大量训练,得到光照变化前后图像像素点之间的映射关系,从而建立了在渐变光照环境下的颜色恒常性模型.该方法不需要内建约束的自适应模型,对于输入的数据不需要对表面属性做特定假设,拥有自适应、自学习的特点.实验结果表明,该模型对室内真实环境中渐变日光下颜色的识别表现出较好的颜色恒常性.     

Neural architecture for robotic vision sensor that discounts uneven illumination in a manufacturing environment

Computer vision algorithms for inspection or pick-and-place operations often depend on spatially uniform illumination of a workplace. This necessitates expensive lighting fixtures. To discount effects of uneven illumination we designed and tested a neural network that can adaptively control light sensitivity at the photosensor level. Our neural network architecture consists of multiple layers with hexagonally arranged nodes. All nodes have partially overlapping receptive fields of different spatial frequencies. Feedforward connections are excitatory while feedback pathways subserve lateral inhibition. The outputs of these nodes are combined so as to maximize the signal-to-noise ratio while constantly resetting thresholds to maintain high sensitivity. Our connectionist architecture can account for many characteristics attributed to the lightness constancy phenomenon observed in biological systems. The results suggest that our module maintains high sensitivity over the whole domain of intensities without interfering with transmission of visual information embedded in spatial discontinuities of intensity.     

基于机器视觉的色差检测算法

针对金属印刷质量中的色差检测问题,采用机器视觉的技术对色差检测算法进行了研究。通过在金属印刷品的留白区域印刷色标,使用工业CCD相机采集金属印刷产品上的色标区域,使用数字图像处理技术提取色标。使用了基于HSV颜色空间的色差检测算法和基于CIELAB颜色空间的色差检测算法,分析了两种颜色空间下色差检测的实验结果,采用两种颜色空间检测算法相结合的方法,实现对色差合理有效的快速检测,同时能保证检测结果的准确性。     

基于视锥细胞色适应生理机制的颜色恒常性计算

通过模拟人类底层视觉系统中的颜色恒常处理机制,提出基于视锥细胞色适应生理机制的颜色恒常性计算方法,利用视锥细胞的色适应计算代替基于灰度边缘假设的颜色恒常性计算方法所采用的Von Kries对角变换.实验结果及定量定性的分析表明本文方法比原计算模型下的算法颜色校正效果要好,能有效恢复图像中物体的本质颜色,且性能更稳定.从计算神经科学的角度,本文方法支持视锥细胞的色适应生理机制在视觉颜色恒常性中扮演重要角色的观点.     

Screen‐based color vision test for digital quantification of deutan color defects using a color‐calibrated monitor

Display technologies are used in various industries. However, few studies on color vision testing using self‐luminous displays are reported. In this study, a screen‐based color vision test, termed the digital color vision test (DCVT), is developed on a color‐calibrated monitor. A control bar is adjusted by the observer to build just noticeable chromatic difference ellipses in CIE u′v′ chromaticity diagram, which indicate the capabilities of subjects' color visions. Ten color‐normal observers and 13 deutan observers participated in the psychophysical experiments. The observers also performed the other two typical diagnostic tests (D‐15 and anomaloscope). Results from the experiment using the adjustment psychophysical method for digital quantification of deutan color vision defects are presented. It is found that the diagnosis accuracy of the DCVT is approximately equivalent to those of D‐15 and anomaloscope tests.     

亮度补偿变换矩阵的颜色恒常性算法

颜色恒常性是计算机视觉的重要研究方向,旨在准确识别目标的真实颜色而不受场景光源变化的影响。目前提出的多种颜色恒常性算法,使用传统的Von Kries 对角变换矩阵对图像的估计照明进行校正,对低照度和强光照射条件下采集到的图像处理效果比较差。根据图像形成的数学模型和光学原理提出了亮度补偿对角变换矩阵的颜色恒常计算方法,该方法对图像颜色校正的同时根据图像像素亮度变化对图像的亮度进行补偿。通过采用多种颜色恒常性算法进行实验验证,该方法能够有效地校正低照度和强光照射图像的颜色、对比度和亮度,从而增强了图像的视见度。     

基于超像素和局部颜色恒常性的自适应阴影去除

为快速有效地去除监控视频中运动目标的投射阴影,提出了一种基于超像素和阴影区域的局部颜色恒常性的自适应阴影去除算法。首先采用改进的简单线性迭代聚类算法将视频图像中的运动前景分割为互不重叠的超像素;然后计算了RGB颜色空间中背景与运动前景的亮度比率,并分析了阴影区域的局部颜色恒常性;在此基础上,以超像素为基本处理单元,计算亮度比率的标准差,并利用阴影区域标准差的特征及其分布规律提出基于拐点的自适应阈值算法检测并去除阴影。实验结果表明,该算法可以适用于多种真实场景下的阴影检测,且阴影检测率与目标识别率均超过85%;基于超像素处理可以大幅度降低算法的计算复杂度,该算法每帧平均处理时间为20 ms。该算法可以同时满足阴影去除对准确度、实时性和鲁棒性的要求。     

Automatic color constancy algorithm selection and combination

In this work, we investigate how illuminant estimation techniques can be improved taking into account intrinsic, low level properties of the images. We show how these properties can be used to drive, given a set of illuminant estimation algorithms, the selection of the best algorithm for a given image. The algorithm selection is made by a decision forest composed of several trees on the basis of the values of a set of heterogeneous features. The features represent the image content in terms of low-level visual properties. The trees are trained to select the algorithm that minimizes the expected error in illuminant estimation. We also designed a combination strategy that estimates the illuminant as a weighted sum of the different algorithms’ estimations. Experimental results on the widely used Ciurea and Funt dataset demonstrate the effectiveness of our approach.     

Multi-objective optimization based color constancy

This paper presents a new combining approach for color constancy, the problem of finding the true color of objects independent of the light illuminating the scene. There are various combining methods in the literature that all of them use weighting approach with either pre-determined static weights for all images or dynamically computed weights for each image. The problem with weighting approach is that due to the inherent characteristics of color constancy methods, finding suitable weights for combination is a difficult and error-prone task. In this paper, a new optimization based combining method is proposed which does not need explicit weight assignment. The proposed method has two phases: first, the best group of color constancy algorithms for the given image is determined and then, some of the algorithms in this group are combined using multi-objective optimization methods. To the best of our knowledge, this is the first time that optimization methods are used in color constancy problem. The proposed method has been evaluated using two benchmark datasets and the experimental results were satisfactory in compare with state of the art algorithms.     

基于视觉颜色感知——光学相似的图像去雾方法

针对雾霾环境影响致使户外获取的图像质量严重下降问题，提出了一种基于视觉颜色感知—光学相似的图像去雾方法。充分利用人眼感知颜色的视觉机理，结合图像的相似性原理，构造了光学相似度函数，建立了新的基于视觉颜色感知—光学相似的图像去雾模型并设计相关算法，进而进行仿真验证。仿真实验结果表明，提出的方法在对有雾图像清晰化处理过程中效果明显，并与现有的图像去雾方法在主观视觉和客观量化方面进行图像去雾效果对比，进一步表明提出的方法在清晰化含雾图像处理中取得了较好的效果。     

A sequential Bayesian approach to color constancy using non-uniform filters

This paper introduces a non-uniform filter formulation into the Brainard and Freeman Bayesian color constancy technique. The formulation comprises sensor measurements taken through a non-uniform filter, of spatially-varying spectral sensitivity, placed on the camera lens. The main goal of this paper is twofold. First, it presents a framework in which sensor measurements obtained through a non-uniform filter can be sequentially incorporated into the Bayesian probabilistic formulation. Second, it shows that such additional measurements obtained reduce the effect of the prior in Bayesian color constancy. For the purposes of testing the proposed framework, we use a filter formulation of two portions of different spectral sensitivities. We show through experiments on real data that improvement in the parameter estimation can be obtained inexpensively by sequentially incorporating additional information obtained from the sensor through the different portions of a filter by Bayesian chaining. We also show that our approach outperforms previous approaches in the literature.     

基于改进图像导数框架的颜色恒常性计算方法

在基于图像导数框架的颜色恒常性算法的基础上,进一步考虑边缘类型和颜色通道间的相关性对于光照色度估计的影响,提出一种基于改进图像导数框架的颜色恒常性计算方法。根据导数图像中像素的饱和度提出一种饱和度权值方案;基于光学特性对导数图像中的边缘进行分类,通过引入照度准不变量提出类型相关的边缘权值方案;将两种权值方案与图像导数框架相结合,并据此进行光照色度估计,利用von Kries对角变换对偏色图像进行校正。实验结果表明该算法有效提高了光照色度估计的准确性,改善了颜色恒常性算法的效果。     

基于机器视觉的采茶机割刀控制方法

针对现有乘坐式采茶机切割的茶叶质量偏低的问题,提出了一种基于机器视觉的采茶机割刀控制方法。首先对相机参数进行低复杂度的标定,然后采用动态阈值分割方法和颜色分类器,提取茶叶图像中的嫩芽区域,并设计间接定位法定位弧形割刀位置;接着通过计算两个指定区域的嫩芽面积及它们的和与差异,得到弧形割刀左右两侧的动作参数;最后将动作参数传给下位机,控制割刀到达预期位置。本方法分别对单株茶树和实地茶园进行了实验,结果表明该方法能够准确定位弧形割刀位置和识别嫩芽区域,实现割刀位置的自适应调整,具有较好的茶叶切割效果,切割得到的嫩芽比例可达70%以上。     

Using eigencolor normalization for illumination-invariant color object recognition

Color is one of salient features for color object recognition, however, the colors of object images sensitively depend on scene illumination. To overcome the lighting dependency problem, a color constancy or color normalization method has to be used. This paper presents a color image normalization method, called eigencolor normalization, which consists of two phases as follows. First, the compacting method, which was originally used for compensating the adverse effect due to shape distortion for 2-D planar objects, is exploited for 3-D color space to make the color distribution less correlated and more compact. Second, the compact color image is further normalized by rotating the histogram to align with the reference axis computed. Consequently, the object colors are transformed into a new color space, called eigencolor space, which reflects the inherent colors of the object and is more invariant to illumination changes. Experimental results show that our eigencolor normalization method is superior to other existing color constancy or color normalization schemes on achieving more accurate color object recognition.     

Estimation of a fluorescent lamp spectral distribution for color image in machine vision

We present a technique to quickly estimate the Illumination Spectral Distribution (ISD) in an image illuminated by a fluorescent lamp. It is assumed that the object colors are a set of colors for which spectral reflectances are available (in our experiments we use spectral measurements of 12 colors checker chart), the sensitivities of the camera sensors are known and the camera response is linear. Thus, the ISD can be approximated by a finite linear combinations of a small number of basis functions.     

结合快速融合与颜色传递的双波段夜视图像染色

总结了近15年来的多传感器图像染色技术研究成果,给出了染色算法的一般模型。在此基础上,将图像融合技术与颜色传递技术结合,设计了基于YCbCr颜色空间的夜视微光与红外图像染色算法。使用互调制快速融合充分提取源图像的空间信息,将得到的融合图像指派给Y分量;借鉴Toet提取图像共同分量的思路,构造Cb、Cr分量;最后应用颜色传递技术,获得了细节丰富、目标背景对比度高,符合场景自然颜色分布的彩色图像。因使用了互调制快速融合方法并直接构造颜色分量,染色过程效率高且参数自适应,可满足实时系统应用需求。     

A sequential machine vision procedure for assessing paper impurities

We present a sequential, two-step procedure based on machine vision for detecting and characterizing impurities in paper. The method is based on a preliminary classification step to differentiate defective paper patches (i.e., with impurities) from non-defective ones (i.e., with no impurities), followed by a thresholding step to separate the impurities from the background. This approach permits to avoid the artifacts which occur when thresholding is applied to paper samples that contain no impurities. We discuss and compare different solutions and methods to implement the procedure and experimentally validate it on a datasets of 11 paper classes. The results show that a marked increase in detection accuracy can be obtained with the two-step procedure in comparison with thresholding alone.