Research on remote sensing image recognition based on BP network information fusion

The many advantages of Neural Network (NN) make it is one of the important technologies for the reorganization of remote sensing images. Taking the remote sensing image from IKONOS satellite as original data and three types of typical targets of ports, airports and bridges as recognition objects, the paper researched on the typical target recognition in the framework of character-level information fusion technology with the algorithm tool of BP NN. The reorganization experiment based on BP NN verifies the feasibility of improving the recognition accuracy of information fusion with spectral characteristics, texture characteristics and geometry structure characteristics.