查询结果:   马成前,王利,刘畅.基于机器学习的图像偏色检测[J].计算机应用与软件,2018,35(3):199 - 204.
中文标题
基于机器学习的图像偏色检测
发表栏目
图像处理与应用
摘要点击数
882
英文标题
COLOR CAST DETECTION BASED ON MACHINE LEARNING
作 者
马成前 王利 刘畅 Ma Chengqian Wang Li Liu Chang
作者单位
武汉理工大学计算机科学与技术系 湖北 武汉 430070     
英文单位
Department of Computer Science and Technology, Wuhan University of Technology, Wuhan 430070,Hubei, China     
关键词
偏色 机器学习 视频图像 灰度 图像分割
Keywords
Color cast Machine learning Video imagery Gray level Image segmentation
基金项目
作者资料
马成前,教授,主研领域:地下空间智能监控系统,系统集成,项目管理。王利,硕士生。刘畅,博士生。 。
文章摘要
图片偏色检测目前大多是利用统计学知识,通过计算各个颜色通道关系判断偏色。目前针对马路摄像机偏色检测的研究工作较少,据此提出一种基于机器学习的马路视频图像偏色检测模型。在YUV色彩空间下,对图像过滤马路的灰色信息,针对部分偏色采用Floyd算法将图片分割。根据色彩分布提取特征值,采用监督性机器学习算法,输入特征向量得到分类器。通过实验表明采用BP神经网络和SVM支持向量机算法,能得到很好的结果。实验结果表明该模型,对多种偏色情况、多种偏色状态有良好的适应性。
Abstract
Most of the current color cast detection by using the knowledge of statistics, through the calculation of whether each color channel. There are few researches on the color cast detection of road cameras, so a color cast detection model of video images based on machine learning was proposed. In the YUV color space, the gray information of the road was filtered and the Floyd algorithm was used to segment the image. Then, the eigenvalues were extracted according to the color distribution, and the supervised machine learning algorithm was used to input the eigenvectors to obtain the classifier. Experiments showed that we got good results by using BP neural network and SVM support vector machine algorithm. The model had good adaptability to many kinds of color castings and various color castings.
下载PDF全文