摘要
对原始直方图统计方法、加入核函数加权的直方图统计方法以及加入背景加权的直方图统计方法进行对比分析,比较核函数加权与背景加权的优缺点。针对背景加权的不足结合图像差分进行改进,一定程度上增强背景加权的鲁棒性。最后通过一系列有针对性的实验,用概率分布图的形式直观的表现各种直方图统计方法的局限性。
Three histogram computation methods are interpreted and compared, including the initial histogram computation method, the kernel weighted histogram computation method, and the background weighted histogram computation method, The merits and demerits of kernel weighted method and background weighted method are compared. The background weighted method is improved by combining image di{{erence method to overcome its demerits, wl^ich increases the robustness of background weighted method to a certain extent. Finally, these methods＇ limitations are showed in virtue of probability distribution images in some special experiments.
出处
《计算技术与自动化》
2007年第4期71-74,共4页
Computing Technology and Automation
基金
国家自然科学基金项目(60775023)
山东省自然科学基金项目(Z2005G03)
关键词
直方图统计方法
核函数加权
背景加权
概率分布图
histogram computation method
kernel weight
background weight
probability distribution image
