remote sensing image change detection based on the combination of pixel-level and object-level analysis

In order to improve the change detection accuracy of the high resolution remote sensing image, a novel framework based on the combination of pixel-level and object-level analysis is proposed. Firstly, the two temporal images are superimposed, and the principal component analysis is performed. Then, it is utilized that the entropy rate segmentation algorithm to segment the first principal component image by changing the number of super-pixels to obtain the multi-layer super-pixel regions with different sizes. At the same time, by analyzing the difference of spectral feature and texture feature on two temporal images, it is used that adaptive PCNN neural network algorithm to make a fusion of the two difference images. Afterwards, the level set (CV) method is used to get the pixel-level change detection results. At last, the change intensity level quantization and decision level fusion are used on the initial change detection results with the region labeling matrix, serving as the post-processing part to obtain the changed objects. Experimental results on the sets of SPOT-5 multi-spectral images show that the new framework can effectively integrate the advantages of pixel-based and object-based image analysis methods, which can further improve the stability and applicability of the change detection process.