Faster-RCNN电力输送塔检测算法  

Target Detection Algorithm for Faster-RCNN Power Transmission Tower

作　　者：严星 尤洪峰 YAN Xing;YOU Hong-feng(School of Computer Science and Engineering,Xinjiang University of Finance and Economics,Urumqi Xinjiang 830012,China;School of Information Science and Engineering,Xinjiang University,Urumqi Xinjiang 830003,China)

机构地区：[1]新疆财经大学计算机科学与工程学院,新疆乌鲁木齐830012 [2]新疆大学信息科学与工程学院,新疆乌鲁木齐830000

出　　处：《计算机仿真》2020年第2期135-139,298,共6页Computer Simulation

基　　金：新疆维吾尔自治区自然科学基金项目(2016D01C050);新疆自治区科技人才培养项目(QN2016YX0051)。

摘　　要：提出了一种基于Faster-RCNN改进的目标检测算法。利用局部旋转和基于参考白算法,对多角度拍摄的无人机遥感影像进行预处理,以减少图片噪声和修改图片光照、色彩偏差;借助图像特征标注法,进一步加强图像中待检测物体关键性特征的利用率。最后,通过区域生成网络(RPN)的最小尺寸自适应和非极大值抑制阈值适应修改能力,解决大型目标局部被当作完整目标检测的问题,实现目标物体的正确检测。结果表明,相对Faster-RCNN算法,检测精度上提高了7.64%。An improved target detection algorithm based on Faster-RCNN is proposed in the paper.The local rotation and the reference white algorithm were used to preprocess the unmanned aerial vehicles(UAV)images of multiple angles to reduce the image noise and modify the picture light and color deviation.With the help of image feature tagging,the utilization rate of the key features to be detected in the image was further strengthened.Finally,the minimum size adaptive and non maximum value suppression threshold of the region generation network(RPN)were used to adapt the modification ability and solve the problem that the large target is partially detected as a complete target detection,and the correct detection of the target object was realized.The results show that the detection accuracy is increased by 7.64% compared with the Faster-RCNN algorithm.

关 键 词：局部旋转 参考白算法 区域生成网络 超参数自适应 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]