改进Mask R-CNN算法在低光道路环境下行人检测研究

Improved Mask R-CNN Algorithm for Pedestrian Detection in Low-Light Road Conditions

针对目标检测算法在低光条件下检测性能下降的问题,以Mask R-CNN目标检测算法为基础,将提出的图像融合模块(MSRCR-IF)引入该目标检测网络中,同时为了更好地利用特征信息,改进了特征金字塔网络,并通过调整区域提交网络以及去除实例分割分支的方式,减少检测目标所花费的时间。实验结果显示:在COCO2017数据集下提出的算法优于其他主流算法,同时在自行构建的低光道路行人数据集下达到了75.05%的平均检测精度,比改进前提高了4.66%。为了验证改进算法的有效性,进行了实车数据测试,结果显示:改进方法能有效提高低光条件下行人检测效果。

Aiming at the problem of the performance degradation of the target detection algorithm in low-light conditions,an image fusion module(MSRCR-IF),proposed in this paper,is introduced into the target detection network,which is based on the Mask R-CNN target detection algorithm.Meanwhile,the feature pyramid network is improved to make better use of feature information,and the region Proposal network is adjusted and the instance segmentation branch is removed to reduce the time of target detection.Experimental results show that the proposed algorithm is better than other mainstream algorithms under the COCO2017 data set.At the same time,it achieves an average detection accuracy of 75.05%under the self-build low-light road pedestrian data set,which is 466%higher than the traditional algorithm.In order to verify the effectiveness of improved algorithm,a real-vehicle data test is carried out.The test results show that the improved method can effectively improve the pedestrian detection effect under low light conditions.