复杂背景下远距离及小尺寸行人检测改进算法  

Improved algorithm for long-distance and small-size pedestrian detection in complex backgrounds

作　　者：谢永明 王红蕾[1] XIE Yong-ming;WANG Hong-lei(College of Electrical Engineering,Guizhou University,Guiyang 550025,China)

机构地区：[1]贵州大学电气工程学院,贵州贵阳550025

出　　处：《计算机工程与设计》2021年第5期1323-1330,共8页Computer Engineering and Design

基　　金：国家自然科学基金项目(61861007);黔科合平台人才基金项目([2016]5103)。

摘　　要：针对远距离及小尺寸行人难以检测,极易受到复杂背景的干扰,且分辨率低、有效信息少等问题,提出一种在复杂背景下结合Faster R-CNN检测远距离及小尺寸行人的改进算法。运用混合高斯模型,解决复杂背景干扰的问题,及时去除视频的背景信息,提取视频图像的前景。为进一步解决分辨率低且有效信息少的问题,采用双线性二次插值方法增强图像的分辨率,采用多尺度特征融合弥补有效信息的不足。实验结果表明,不同场景下行人检测精度均有所提升,其中远距离及小尺寸行人的检测精度提升更为明显。Long-distance and small-sized pedestrians are difficult to detect,and they are easily interfered by complex backgrounds.In addition to the problems of low resolution and less effective information,an improved method combining Faster R-CNN to detect long-distance and small-sized pedestrians in a complex background was proposed.The mixed Gaussian model was used to solve the problem of complex background interference.The background information of the video was removed in time and the foreground of the video image was extracted.To further solve the problems of low resolution and less effective information,the bilinear quadratic interpolation method was used to enhance the resolution of the image,and multi-scale feature fusion was used to make up for the lack of effective information.Experimental results show that the detection accuracy of pedestrians in different scenarios is improved,and the detection accuracy of long-distance and small-sized pedestrians is improved significantly.

关 键 词：卷积神经网络 行人检测 混合高斯模型 双线性二次插值 多尺度特征融合 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]