A method for segmentation and extraction of cultivated land plots from high-resolution remote sensing images

Aiming at the problem of segmentation and extraction of cropland parcels in remote sensing images with complex background, a method of segmentation and extraction of cropland in high-resolution remote sensing images is proposed. Firstly, the linear spectral clustering (LSC) algorithm is applied to the remote sensing image to obtain the segmentation results of superpixel blocks in the target area and the background area; then, the maximum similarity region merging algorithm (MSRM) algorithm is used to merge the superpixel blocks of two different areas separately, which effectively reduces the phenomenon of under-segmentation and over-segmentation of the image and obtains the binary image containing the cultivated land parcels and non-cultivated land parcels. Based on this, the total arable area is calculated using MATLAB. Finally, in order to verify the correctness and effectiveness of the proposed method, the remote sensing image data provided by Beijing Guosheng Xingmai Information Technology Co. The simulation results show that the proposed method can effectively segment and extract remote sensing cropland images.