Heterogeneous Change Detection of Optical and SAR Based on Mixed Attention Mechanism

Heterogeneous change detection can combine the advantages of multi-source images and capture change information more comprehensively. Firstly, the log-ratio operator and Hierarchical Fuzzy C-Means (HFCM) algorithm are used to suppress speckle noise and enhance the contrast between different regions to generate the initial change map. Then, the convolutional neural network is used as the main framework and the convolutional block attention module (CBAM) attention mechanism is combined to make the network focus on the key regions, improve the ability of model to recognize the changing regions, and obtain the final detection result. Experimental results on heterogeneous data sets of SAR and optical images demonstrate that the proposed method can achieve higher accuracy and reliability.