Remote Sensing Image Recognition Method Based on Faster R-CNN

This paper proposes a method for remote sensing image recognition based on Faster R-CNN. Using Faster R-CNN model and ZFNet as the basic network, experiments show that the accuracy rate of Architecture, Greenhouses and Paddy field recognition is 90.67%, 93.85%, 83.33%, and the average recognition accuracy reached 89.28%. At the same time, compared with the recognition results of recognition detection methods such as CNN and TT-RICNN, it was found that the proposed Faster R-CNN model has better recognition performance well, with good recognition detection accuracy.