机电工程技术2025，Vol.54Issue(6)：140-143,4.DOI:10.3969/j.issn.1009-9492.2025.06.024
基于SAR图像区域约束的输电杆塔智能识别方法
Power Transmission Towers Detetion Based on SAR Images and Deep Learning
摘要
Abstract
Satellite-based synthetic aperture radar(SAR),as an active microwave imaging technology,provides all-weather,all-time earth observation and is increasingly used in remote identification and monitoring tasks for infrastructure.Address the issues of false alarms and missed detections in the detection of transmission towers in high-resolution SAR images in complex scenarios,a dual-layer object detection method combining regional constraints is proposed.Firstly,a SAR dataset of transmission towers from different regions is constructed,including target-level and regional-level annotations.Based on this,deep learning models such as YOLOv8 are utilized to recognize transmission towers,and the detection accuracy,inference speed,and applicability of different models in large-area SAR images are compared.The results show that the proposed method significantly reduces false alarm occurrences in urban and water areas.In test cases,the recognition rate of transmission towers reaches 90%,demonstrating the effectiveness of the proposed method.关键词
SAR/输电杆塔/深度学习/YOLOv8Key words
SAR/power transmission tower/deep learning/YOLOv8分类
信息技术与安全科学引用本文复制引用
岑宗浩,范鹏,刘波..基于SAR图像区域约束的输电杆塔智能识别方法[J].机电工程技术,2025,54(6):140-143,4.基金项目
国家电网有限公司华东分部项目(52992424001X) （52992424001X）
