Ground feature identification method and system for power transmission line engineering survey task

• Main Authors: LI JIANDONG, HU NAN, BEN SHUJUN, SEO JUNUL, YAN YULI, YUAN BEI'ER, YANG PENGPENG, LI NA, HU DONGKANG, CHENG YANTAO, DENG DA, ZHANG XIAOXIANG, HUA LEI
• Format: Patent
• Language: Chinese; English
• Published: 23.07.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention belongs to the technical field of image processing, and particularly relates to a ground feature identification method and system for a power transmission line engineering survey task. According to the method, a high-resolution inclined aerial image is used as a data source, and a basic sample data set of typical ground feature elements such as houses, roads and trees is generated by using an integrated large model algorithm labeling tool. Meanwhile, a basic instance segmentation network (HRFI-YOLO network) suitable for surface feature recognition of a high-resolution inclined aerial image is designed, an aerial image surface feature recognition model is generated through training, and the model can extract surface feature targets such as houses, roads and trees in a remote sensing image. According to the method, the availability of the label data and the stability of the deep learning model are ensured while the workload is reduced. By constructing the deep learning network model, automatic ext