Local Climate Zone Classification Using Daytime Zhuhai-1 Hyperspectral Imagery and Nighttime Light Data

The tremendous advancement of cities has caused changes to the urban subsurface. Urban climate problems have become increasingly prominent, especially with regard intensification heat island (UHI) effect. local zone (LCZ) is a new quantitative method for analyzing that based on kind surface and can effectively deal problem hazy distinction between rural areas in UHI effect research. LCZs are widely used regional modeling, planning, thermal comfort surveys. Existing large-scale LCZ classification methods usually use visual features optical images, such as spectral textural features. There many hyperspectral extraction over large areas. an integrated concept includes geography, society, economy. Consequently, it makes sense consider characteristics human activity images interpret them accurately. ALOS_DEM data depict city’s physical characteristics; however, nighttime lights crucial indicators activity. These three datasets be combination portray environment. Therefore, this study proposes fusing daytime mapping, i.e., Zhuhai-1 their derived feature indices, data, light from Luojia-1. By combining information, proposed approach captures temporal dynamics areas, providing more complete representation characteristics. integration allows refined identification characterization land cover. It comprehensively integrates exploits synergistic information multiple sources, provides higher accuracy resolution mapping. First, we extracted various features, namely spectral, red-edge, Random forest (RF) XGBoost classifiers were used, average impurity reduction was employed assess significance variables. All input variables optimized select best results 5th ring road area Beijing, China, revealed technique achieved mapping good precision, total 87.34%. In addition, examine contrast effects indices accuracy, used. showed accuracies terms improved by 2.33% 2.19% using RF classifier, respectively. radiation brightness value (RBV) (GI = 0.0212) attained classification’s highest variable importance value; DEM also produced high GI (0.0159), indicating night lighting landform strongly influence classification.