Neighborhood Extremum Method of Extracting Urban Built-Up Area Using Nighttime Lighting Data

To address the problem of low accuracy in the urban built-up area extraction method using nighttime light data due to light spillover characteristics, the build-up area extraction method based on neighborhood extremum is proposed. Firstly, the one-dimensional quadratic regression is used to perform relative radiation correction for nighttime light data. Then, the extremum images describing the spatial variation characteristics of gray values are obtained by extremum neighborhood filtering. Finally, the extremum search algorithm was used to obtain the boundary images of built-up areas, and the binary segmentation method is used to extract urban built-up areas. The experimental results show that the means of Kappa coefficients and threshold selection times of our proposed method are 0.85 and 37 s, which are 0.03, 1 503 s and 0.01, 443 s higher than that of the mutation detection method and the statistical analysis method. The spatial morphology of the built-up area extraction results is closer to the reference data, which has better extraction effect and stability.