基于卫星图像的道路检测系统的设计与实现  

Design and Realization of Road Detection System Based on Satellite Image

作　　者：康长青[1] 杭波[1] 周哲[1] 杨明 朱丽娟[1] KANG Changqing;HANG Bo;ZHOU Zhe;YANG Ming;ZHU Lijuan(School of Computer Engineering,Hubei University of Arts and Science,Xiangyang,Hubei,441053,China;Xiangyang Yuqing Transmission Technology Co.,Ltd,Xiangyang,Hubei,441004,China)

机构地区：[1]湖北文理学院,湖北襄阳441053 [2]襄阳宇清电驱动科技有限公司,湖北襄阳441004

出　　处：《长江信息通信》2022年第7期69-71,共3页Changjiang Information & Communications

基　　金：湖北省教育厅科学研究计划重点项目(D20182603)。

摘　　要：针对传统的卫星图像道路检测方法存在的手工选取特征困难,计算量大,效果不好的缺点,提出了基于卷积神经网络的卫星图像道路检测系统,主要包括卫星图像预处理、特征提取、模型构建、决策分类和性能评价5个功能模块。系统首先采用中值滤波进行图像预处理,然后利用Canny算子进行特征提取,接着基于卷积神经网络利用数据集进行模型生成,最后利用训练好的检测模型进行道路检测和性能评价。仿真实验表明,系统可以达到77.8%的正确性和77.4%的完整性。The traditional satellite image road detection algorithms have the disadvantages of difficulty in manual feature selection, large amount of calculation and poor performance. A road detection system in satellite images based on convolutional neural network is proposed to solve the above problem. The proposed system mainly includes five functional modules including satellite image preprocessing, feature extraction, model construction, decision classification and performance evaluation.The proposed system first uses median filtering for image preprocessing, then uses Canny operator for feature extraction, then uses data sets to generate models based on convolutional neural networks, and finally uses the trained detection model for road detection and performance evaluation.Simulation experiments show that the system can achieve 77.8% correctness and 77.4% completeness.

关 键 词：道路检测 卫星图像 特征提取 中值滤波 CANNY算子 卷积神经网络 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]