Research On Change Detection Of Urban Land Usage Based On High Resolution Remote Sensing Images

With the development of the national economy,China’s urbanization process is also accelerating,and research on the changes of urban land use has become the focus of current attention.This paper proposes a block change detection method based on multi-scale high-resolution remote sensing image multi-scale segmentation classification results fusion,which can make an intuitive display of the degree of urban land use change.In this paper,through the research and thinking on the traditional change detection algorithm,the proposed urban land use urban change detection method focuses on multi-scale image segmentation multi-parameter selection,classification result fusion,and multi-temporal image block change detection.The accuracy of the classification results directly affects the accuracy of the change detection results.Compared with the conventional classification method based on the optimal segmentation scale,this method classifies the multi-scale segmentation results separately and fuses the classification results to improve the overall accuracy..Through the accuracy analysis of the experimental results of two different images,the overall classification accuracy of the method is improved compared with the classification accuracy under the various scales before the fusion.Traditional post-classification change detection is often done by ArcGIS through correlation analysis.This paper will classify the classification results,and use the block as the research unit to calculate the difference by calculating the difference of the small square histogram,taking into account the small block.Structural features,the small blocks continue to be divided into small units and obtain histograms,combined to form a new histogram;this paper measures the difference between the two histograms by the Pap address and the pure increment percentage of the block change,and the resulting graph can contain The change detection heat map and the change plot map after the binary splitting can achieve the effect of similar change type screening by corresponding condition setting.At the same time,the operation result of the algorithm is stored in the form of a cell matrix,and the extraction of different change types only needs to be searched from the result matrix.The experimental results show that the overall accuracy of the change detection algorithm is high.