为了更准确地得到符合车辆形态学特征的初始候选框，提出一种基于改进的FasterR-CNN模型的车辆检测算法.首先提取目标框的人工标注坐标值，得到标注框的宽度和高度，然后利用K-means算法对所有框的宽高值进行聚类，得出聚类中心点坐标值后，重新设置RPN的锚盒尺寸及比例，对Faster R-CNN算法的3种尺寸和3种比例加以改进.最后选择轿车、SUV、客车和货车4种车型车辆数据，对改进前后的Faster R-CNN模型进行训练，比较改进前后的模型在车辆检测及车型识别任务中的表现.实验结果表明，使用改进的FasterR-CNN模型达到86.54%的检测准确率，较原始模型提高3.12%.并且该模型有效解决了漏检和误检问题，在恶劣天气和复杂背景下均表现出较高的鲁棒性.
In order to obtain an initial candidate frame that conforms to the morphological characteristics of vehicles more accurately, a vehicle detection algorithm based on the improved Faster R-CNN model is proposed. First, the coordinate values of target frames are extracted to get width and height values of labeled boxes, and then K-means algorithm is used to cluster the width and height values of all boxes. Second, by resetting the anchor box size and the anchor box ratio of region proposal network (RPN) according to the coordinates of the cluster center point, the three sizes and three ratios of the Faster R-CNN can be improved. Finally, vehicle data of four types including cars, SUVs, buses and trucks are selected to train both the unimproved and the improved Faster R-CNN models. At the same time, the performance of the two models in vehicle detection and vehicle identification tasks are compared. Experimental results show that the improved Faster R-CNN model can achieve 84.69% detection accuracy, which is 3.12% higher than the original model. The algorithm effectively improves the missed detection and false detection problems, and shows high robustness in bad weather and complex background.
[1] 刘学亮,严捷丰,周荷琴.基于轮廓特征的车辆遮挡检测和分离算法[J].计算机仿真,2009, 26(4):297-300. Liu X L, Yan J F, Zhou H Q. Vehicle occlusion detection and separation algorithm based on contour features[J]. Computer Simulation, 2009, 26(4):297-300.(in Chinese)
[2] 黄银花,赵仕奇.基于特征模型驱动的前方车辆检测[J].计算机测量与控制,2007(11):1478-1479, 1502. Huang Y H, Zhao S Q. Front vehicle detection based on feature model driving[J]. Computer Measurement & Control, 2007(11):1478-1479, 1502.(in Chinese)
[3] 郭磊,李克强,王建强,等.一种基于特征的车辆检测方法[J].汽车工程,2006(11):1031-1035. Guo L, Li K Q, Wang J Q, et al. Vehicle detection method based on features[J]. Automotive Engineering, 2006(11):1031-1035.(in Chinese)
[4] 刘勃,周荷琴.昼夜转换场景中的车辆检测[J].信号处理,2006(3):390-394. Liu B, Zhou H Q. Vehicle detection in day-night transition scenarios[J]. Journal of Signal Processing, 2006(3):390-394.(in Chinese)
[5] 付彦君,裘伟,戴斌,等.基于多特征融合的车辆实时检测[J].计算机仿真,2007(10):212-215. Fu Y J, Qiu W, Dai B, et al. Real-time vehicle detection based on multi-feature fusion[J]. Computer Simulation, 2007(10):212-215.(in Chinese)
[6] 蔡英凤,刘泽,孙晓强,等.基于双车辆可变形部件模型的车辆检测方法[J].汽车工程,2017, 39(6):710-715, 721. Cai Y F, Liu Z, Sun X Q, et al. Vehicle detection method based on two-vehicle deformable component model[J]. Automotive Engineering, 2017, 39(6):710-715, 721.(in Chinese)
[7] 陈远,谢昭,吴克伟.面向部件遮挡补偿的车辆检测模型[J].中国图象图形学报,2014, 19(12):1802-1811. Chen Y, Xie Z, Wu K W. Vehicle detection model for component occlusion compensation[J]. Journal of Image and Graphics, 2014, 19(12):1802-1811.(in Chinese)
[8] Girshick R, Donahue J, Darrell T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Columbus, OH, USA, 2014:580-587.
[9] Girshick R. Fast R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision, Santiago, Chile, 2015:1440-1448.
[10] Ren S, He K, Girshick R, et al. Faster R-CNN:towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2017, 39(6):1137-1149.
[11] 冯小雨,梅卫,胡大帅.基于改进Faster R-CNN的空中目标检测[J].光学学报,2018, 38(6):250-258. Feng X Y, Mei W, Hu D S. Aerial target detection based on improved Faster R-CNN[J]. Acta Optica Sinica, 2018, 38(6):250-258.(in Chinese)
[12] Li J, Zhang D, Zhang J, et al. Facial expression recognition with Faster R-CNN[J]. Procedia Computer Science, 2017, 107:135-140.
[13] 黄海燕,刘晓明,孙华勇,等.聚类分析算法在不确定性决策中的应用[J].计算机科学,2019, 46(S1):593-597. Huang H Y, Liu X M, Sun H Y, et al. Application of cluster analysis algorithm in uncertainty decision[J]. Computer Science, 2019, 46(S1):593-597.(in Chinese)
[14] He K, Zhang X, Ren S, et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, USA, 2016:770-778.
[15] 常海涛,苟军年,李晓梅. Faster R-CNN在工业CT图像缺陷检测中的应用[J].中国图象图形学报,2018, 23(7):1061-1071. Chang H T, Xun J N, Li X M. Application of faster R-CNN in defect detection of industrial CT images[J]. Journal of Image and Graphics, 2018, 23(7):1061-1071.(in Chinese)
[16] 侯媛媛,何儒汉,李敏,等.结合卷积神经网络多层特征融合和K-Means聚类的服装图像检索方法[J].计算机科学,2019, 46(S1):215-221. Hou Y Y, He R H, Li M, et al. Clothing image retrieval method combining multi-layer feature fusion of convolutional neural network and K-Means clustering[J]. Computer Science, 2019, 46(S1):215-221.(in Chinese)
