A Novel Sample Selection Method for Impervious Surface Area Mapping Using JL1-3B Nighttime Light and Sentinel-2 Imagery

Abstract

Urbanization has attracted wide and active interests due to the impact on regional sustainable development. As an important indicator of urbanization, impervious surface area (ISA) should be accurately monitored. In scenario of identifying ISA by supervised classification from satellite images, the training samples are usually labeled manually, which is highly laborintensive and time-consuming. High-resolution nighttime light image provides a unique footprint of human activities and settlements which are strongly correlated with ISA. In view of this, a novel ISA training sample selection method is proposed by integrating the JL1-3B high-resolution nighttime light imagery and Sentinel-2 time series imagery, and the random forest is applied to classify ISA from Sentinel-2 imagery. The quality of the automatically selected samples was quantitatively validated. Over three study areas, and the overall classification accuracies were above 97%, showing reliable and robust performance. Compared with conventional methods, the proposed approach achieves satisfactory results in separating bare land from ISA. This study provides a data fusion way which can automatically generate sufficient and high-quality training samples for ISA mapping, and suggests that high-resolution nighttime imagery could demonstrate a promising potential for urban remote sensing.

A Novel Bayes Approach to Impervious Surface Extraction from High-Resolution Remote Sensing Images

“…Currently, remote sensing technology is an effective tool for obtaining impervious surface information [ 11 ]. Remote sensing technology can provide accurate spatial and temporal information on the Earth’s surface by allowing simultaneous observation of large areas in a short period [ 12 ]. With the development of satellite images in recent years, various low-, medium- and high-resolution satellite images have been widely used for impervious surface extraction studies [ 13 ].…”

“…Many global and regional land cover products that can be used for impervious surface information analysis have been released in recent years through a supervised classification framework [ 43 , 44 ]. However, supervised classification produces products with low spatial resolution and limited local accuracy [ 12 ]. In summary, medium-low-resolution impervious surface mapping techniques have matured, but relatively few studies have been conducted on high-resolution impervious surface information extraction [ 16 ].…”

A Comparative Analysis of Index-Based Methods for Impervious Surface Mapping Using Multiseasonal Sentinel-2 Satellite Data

“…Only the performance of NDBI, however, was assessed using Sentinel-2 images. Tang et al [37] developed a new ISA extraction method using JL1-3B Nighttime Light and Sentinel-2 Image. In their study, NDBI and BCI were used to evaluate the extraction performance.…”

“…To generate an impervious surface thematic map for spatial and quantitative analysis, the calculated index images need to be reclassified as either impervious or pervious. Commonly used methods include thresholding methods [18], [37], [38] and unsupervised classification methods [39], [40]. It is widely acknowledged that the selection of binary algorithms may have an influence on the accuracy of ISA extraction [38], [41].…”

Classificação De Superfícies Impermeáveis Em Imagem Multiespectral Com Algoritmo De Machine Learning

“…Kesikoglu et al, 2021 mostraram resultados promissores na identificação de áreas impermeáveis com a utilização do algoritmo Support Vector Machine (SVM) e imagens Landsat. Tang et al, (2020) A aprendizagem de máquina (Machine Learning -ML) corresponde a uma subárea da Inteligência Artificial. As técnicas de ML consistem na aplicação de um ou mais algoritmos para a realização de uma determinada tarefa e podem ser aplicados à diversas imagens de alta resolução (MISRA et al, 2020).…”