雨天图像中目标显著性检测算法设计与效果评价

雨天条件下,图像中目标的许多特征被掩盖,使得户外图像应用系统效能发挥受到严重的影响。为了提高雨天条件下图像中目标检测的质量,通过综合分析雨天图像中的目标特征,发现其亮度颜色信息、色彩差异信息和暗通道先验信息对目标的显著性具有高敏感度,进而提取了雨天图像中目标的显著性特征,构建了基于混合特征的目标显著性检测模型,最后通过多个评价指标的效能评估实验,与经典算法进行对比,验证了本文算法的有效性。

Design and Effect Evaluation of Target Saliency Detection Algorithm for Objects in Rainy Day Images

Under rainy conditions, many features of the object in the image are concealed, which seriously affects the performance of outdoor image application system. In order to improve the quality of target detection in images under rainy conditions, the brightness color information, color difference information and dark channel prior information are found to be highly sensitive to the saliency of objects through comprehensive analysis of object features in rainy day images. Then, the salient features of the object in the rainy day image are extracted, and a target salient detection model based on mixed features is constructed. Finally, the effectiveness of the proposed algorithm is verified by comparing with the classical algorithm through the effectiveness evaluation experiments of several evaluation indexes.