基于车载计算机的红外图像移动目标检测

doi:10.12382/bgxb.2022.A012

摘要/Abstract

摘要： 环境感知系统作为无人车平台的重要组成部分，是路径规划与决策控制等功能实现的基础与前提。结合车载计算机的硬件基础，设计一种复杂场景下的红外图像移动目标检测系统。该系统采用客户端/服务端(B/S)架构，在客户端浏览器进行服务请求后，服务端使用改进的YOLOv2算法对车载Rapid输入输出(RapidIO)高速总线传输来的红外图像进行目标检测；通过维度重聚类和改进激活函数等方法，有效降低了漏检和误检率；为弥补红外图像对比度低、纹理特征弱的缺点，提出一种基于直方图均衡化的双边滤波图像增强算法，该预处理能够有效保持图像轮廓细节信息。结果表明：改进YOLOv2算法的实时检测准确度相对于原始算法提升了4.1%；图像预处理算法显著提高了检测准确率；该系统可为开拓无人车载红外图像的应用领域提供可靠的技术支撑。

关键词: 红外图像, YOLOv2算法, 车载计算机, 移动目标检测

Abstract: As an important part of driverless vehicle platform，the environmental perception system is the basis and premise to realize as the path planning and decision control function. Based on the onboard computer，an infrared image moving object detection system in complex scenes is designed.The browser/server architecture is used in the detection system. After the client browser requests for services，the server applies the improved YOLOv2 algorithm to detect the infrared images transmitted by the onboard RapidIO high-speed bus in real time.the missed detection and error detection rates are effectively reduced by dimensional re-clustering and improved activation function.In order to make up for disadvantages of low contrast and weak texture features of infrared images，bilateral filtering image enhancement algorithm based on histogram equalization is proposed，which can effectively maintain the image contour details.The results show that the real-time detection accuracy of the improved YOLOv2 is increased by 4.1% compared with the original algrithm and the image preprocessing significantly improves the detection accuracy.

Key words: infraredimage, imagedenoisingandenhancing, YOLOv2algrithm, onboardcomputer, movingobjectdetection

中图分类号:

TP391.4

李博，王博，韩京冶，杨宗睿，栗霖雲，傅畅之. 基于车载计算机的红外图像移动目标检测[J]. 兵工学报, 2022, 43(S1): 66-73.

LI Bo， WANG Bo， HAN Jingye， YANG Zongrui， LI Linyun， FU Changzhi. Infrared Image Moving Object Detection Technology Based on Onboard Computer[J]. Acta Armamentarii, 2022, 43(S1): 66-73.

参考文献

［1］ KIESER D.Driverless vehicles and their impact on your portfolio［J］.Equity，2018，32(1):4.
［2］高凯珺，孙韶媛，姚广顺，等.基于深度学习的无人车夜视图像语义分割［J］.应用光学，2017，38(3):421-428.
GAO K J，SUN S Y，YAO G S，et al.Semantic segmentation of night vision images for unmanned vehicles based on deep learning［J］. Journal of Applied Optics，2017，38(3):421-428. (in Chinese)
［3］冯相琨.基于深度学习的小样本红外图像目标检测与识别研究［D］.北京：北京交通大学，2020.
FENG X K.Research on object detection and recognition of small sample infrared image based on deep learning［D］.Beijing：Beijing Jiaotong University，2020.(in Chinese)
［4］何小亚，王明，张蕾，等.一种双余度车载核心处理机的设计与实现［J］.电子技术，2014(1):29-32.
HE X Y，WANG M，ZHANG L，et al.Design and implementation of a dual redundancy vehicle core processor［J］.Electronic Technology，2014(1):29-32.(in Chinese)
［5］张文俊.一种RapidIO和FC的互联通信技术［J］.电子技术与软件工程，2019(14):18-20.
ZHANG W J.An interconnection communication technology between RapidIO and FC［J］.Electronic Technology and Software Engineering，2019(14):18-20. (in Chinese)
［6］千应庆，王晓锋，劳力，等.一种基于RapidIO协议的光纤总线硬件架构设计与分析［J］.兵工学报，2012，33(12):1480-1484.
QIAN Y Q，WANG X F，LAO L，et al. Hardware architecture design and analysis of a fiber optic bus based on the RapidIO protocol［J］.Acta Armamentarii，2012，33(12):1480-1484. (in Chinese)
［7］ ZHANG B H，JIAO D D，PEI H Q，et al. Infrared moving object detection based on local saliency and sparse representation［J］.Infrared Physics & Technology，2017，86:187-193.
［8］ PINIARSKI K，PAWLOWSKI P.Efficient pedestrian detection with enhanced object segmentation in far IR night vision［C］∥Proceedings of 2017 Signal Processing:Algorithms，Architectures，Arrangements，and Applications.Poznan，Poland:IEEE，2017.
［9］唐诗.基于车载视频的道路车辆及行人检测［D］.成都：电子科技大学，2018.
TANG S.Detection of road vehicle and pedestrian based on on-board video［D］.Chengdu：University of Electronic Science and Technology of China，2018.(in Chinese)
［10］ EVERINGHAM M，GOOL L V，WILLIAMS C K I，et al.The pascal visual object classes(VOC) challenge［J］.International Journal of Computer Vision，2010，88(2):303-338.
［11］ KIM S，SONG W J，KIM S H. Infrared variation optimized deep convolutional neural network for robust automatic ground target recognition［C］∥Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops.Honolulu，HI，US: IEEE，2017:1-8.
［12］熊斌，黄心汉，王敏.基于自适应目标图像恢复的红外弱小目标检测［J］.华中科技大学学报(自然科学版)，2017，45(10) ：25-30.
XIONG B，HUANG X H，WANG M.Infrared dim small target detection based on adaptive target image recovery［J］.Journal of Huazhong University of Science and Technology(Nature Science Edition)，2017，45(10)：25-30.(in Chinese)
［13］ WU Z H，XIAO X H.Study on histogram equalization［C］∥Proceedings of the International Symposium on Intelligence Information Processing and Trusted Computing.Wuhan，China，IEEE，2011.
［14］凡遵林，毕笃彦，马时平，等.亮度保持和细节增强的红外图像增强方法［J］.中南大学学报(自然科学版)，2016，47(6)：1967-1972.
FAN Z L，BI D Y，MA S P，et al.Method for infrared image with brightness preservation and detail enhancement［J］.Journal of Central South University(Natural and Technology)，2016，47(6)： 1967-1972.(in Chinese)
［15］ REDMON J，FARHADI A.YOLO 9000:better，faster，stronger［C］∥Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.Honolulu，HI，US:IEEE，2017:6517-6525.
［16］李珣，刘瑶，李鹏飞，等. 基于Darknet框架下YOLOv2算法的车辆多目标检测方法［J］.交通运输工程学报，2018，18(6):142-158.
LI X，LIU Y，LI P F，et al.Vehicle multi-target detection method based on YOLOv2 algorithm under darknet framework［J］.Journal of Traffic and Transportation Engineering，2018，18(6):142-158. (in Chinese)

[1]	梁杰，李磊，任君，齐航，周红丽. 基于深度学习的红外图像遮挡干扰检测方法[J]. 兵工学报, 2019, 40(7): 1401-1410.
[2]	仇荣超，吕俊伟，宫剑，修炳楠，马新星，刘思彤. 前视红外图像中海岸线与海天线的通用检测方法研究[J]. 兵工学报, 2019, 40(6): 1171-1178.
[3]	刘德鹏，李正周，曾靖杰，熊伟奇，亓波. 基于多尺度局部对比度和多尺度梯度一致性的红外小弱目标检测算法[J]. 兵工学报, 2018, 39(8): 1526-1535.
[4]	王忠美，杨晓梅, 顾行发. 基于鲁棒主成分分析的红外图像小目标检测[J]. 兵工学报, 2016, 37(9): 1753-1760.
[5]	李成，李建勋，童中翔，贾林通，张志波. 红外成像制导末端局部图像识别跟踪研究[J]. 兵工学报, 2015, 36(7): 1213-1221.
[6]	王珺, 彭进业, 何贵青, 冯晓毅, 阎昆. 基于非下采样Contourlet 变换和稀疏表示的红外与可见光图像融合方法[J]. 兵工学报, 2013, 34(7): 815-820.
[7]	李建勋，童中翔，王超哲，柴栋，李贺，张志波. 飞机目标红外特性计算与图像仿真[J]. 兵工学报, 2012, 33(11): 1310-1318.
[8]	于劲松，万九卿，高秀林. 红外图像弱小点目标检测技术研究[J]. 兵工学报, 2008, 29(12): 1518-1521.
[9]	周培德，付梦印，黄源水，刘羿彤. 紅外图像边缘提取的算法[J]. 兵工学报, 2007, 28(5): 524-527.
[10]	王世允，常本康，张俊举，钱芸生，孙恋君，喻春雨. 非制冷热瞄具图像实时开窗增强方法研究[J]. 兵工学报, 2007, 28(2): 163-166.
[11]	周培德，付梦印，张长江. 红外图像边缘提取的随机阈值方法[J]. 兵工学报, 2005, 26(2): 177-180.
[12]	1:王宏波，庄志洪，张清泰,贺歆;2:郑华利. 终端红外图像制导信息处理技术研究[J]. 兵工学报, 2004, 25(6): 689-693.
[13]	1:李宏贵，李兴国;2:罗正发，李国桢. ー种基于遗传算法的紅外图像识别方法[J]. 兵工学报, 2000, 21(1): 19-22.