夜间灯光融合多源数据的建成区提取方法

A Method of Built-up Area Extraction Based on Nighttime Light Fusing Multi Source Data

夜间灯光数据已被广泛应用于城市建成区提取研究。针对已有研究未过滤NPP/VIIRS数据火光、气体燃烧等背景噪声,且仅顾及单一要素进行建成区提取,导致建成区提取不准确的问题,提出一种夜间灯光融合多源数据的建成区提取方法。首先,构建顾及POI、道路、EVI多源要素的夜间灯光修正指数PREANI;然后,基于该指数,利用动态阈值二分法对建成区进行提取;最后,以东营市为实验区,对提出的方法进行对比验证和讨论。结果表明:该方法在城市边缘地物区分、城市空间结构辨识能力方面具有明显优势,对于整体空间形态、细节边界表现效果更优;基于该方法提取的建成区,查全率、查准率和F1值均有明显提高,其提取边界更为精准,表现出较好的适用性。本研究对于城市建成区提取与城市空间结构刻画具有重要的意义。

Nighttime light data has been widely used in urban built-up area extraction.Aiming at the problem of unfiltered background noise such as fire and gas combustion from NPP/VIIRS data,and considering only a single element to extract built-up area,this paper proposes a built-up area extraction method based on nighttime light fusing multi-source data.Firstly,construct the nighttime light correction index PREANI,which takes multi-source elements of POI,road and EVI into account.Then,extract the built-up area by using the dynamic threshold dichotomy method based on the index.Finally,take Dongying city as the experimental area to verify the proposed method.The results indicate that:the built-up area extraction method proposed in this paper has obvious advantages in urban edge feature identification and urban spatial structure identification,and it has better performance for overall spatial morphology and detailed boundary;based on the built-up area extracted by this method,the recall,precision and F1 values are significantly improved,and the extraction boundary is more accurate,showing its good applicability.The research is of great significance for the extraction of urban built-up areas and the characterization of urban spatial structure.