摘要
提出了一种基于颜色特征提取和图像边缘检测融合的运动目标检测算法。利用颜色自相关矩阵对图像建立特征模型,进行相似度的比较得到较粗糙的目标检测结果。然后通过Canny算子提取图像的边缘信息并利用二值差化求得目标大致边缘轮廓。将两者再进行逻辑与操作,即判定在粗糙检测结果中运动目标轮廓内的元素目标边缘信息为运动目标边缘,并最终进行形态学操作得到运动目标。实验结果表明,提出的算法对复杂背景下运动目标有较为理想的检测效果。
The paper proposes an algorithm for the detection of motion objects based on the integration of color features extraction and image edge detection. The color auto-correlogram is used for modeling the features for the images,and the similarities of the detection are compared to obtain the coarse results. Then the edge information is extracted through the Canny edge detection operator with the difference of the consecutive frames binarized to obtain the coarse edge results of the objects. The final results stand out after the morphology operation. The experiments indicate the efficiency of the algorithm for the motion detection in complex environments.
出处
《电子科技》
2014年第9期8-12,共5页
Electronic Science and Technology
基金
国家自然科学基金资助项目(61371175)
黑龙江博士后研究发展基金资助项目(LBH-Q09128)
