考虑光照影响的静态图像感兴趣区域提取  

Approach to Extracting Region of Interests for Static Images Considering Factors of Illumination

作　　者：樊万梅[1] 孙棣华[1] 赵敏[1] 张路[1] 

机构地区：[1]重庆大学自动化学院,重庆400030

出　　处：《世界科技研究与发展》2012年第2期280-283,共4页World Sci-Tech R&D

基　　金：国家高技术研究发展计划(511-0910-1031);国家教育部博士点专项基金(20090191110022)资助

摘　　要：考虑光照对感兴趣区域提取的影响,提出了一种新的基于边缘的感兴趣区域提取算法。该算法基于二维离散小波变换去除了光照对感兴趣区域的影响;并通过对所采集的边缘像素点评价进一步去除了光照在感兴趣区域产生的阴影边缘。将该算法运用于长途客车过道区域的提取。实验结果表明,提出的算法不仅能实现多种情况下的感兴趣区域的自动准确的提取,还能避免光照的影响。Considering the illumination effects,a new region of interest （ROI） extraction algorithm based on edge information is proposed. Based on two-dimensional discrete wavelet transform, the algorithm removes the effects of illumination, then through the acquisition of the edge pixel evaluation further removes the edge produced by the illumination. The region of interest extraction algorithm is employed to extract the corridor region. The results show that the algorithm is efficient in extracting ROI in variety of situations, and avoid the illumination effects.

关 键 词：感兴趣区域 小波变换 评价方法 过道区域 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]