Identification and Extraction of Camellia Oleifera Based on Multi-source Data

Taking Liaoshi town, Yuanzhou district, Yichun city, Jiangxi province, as the study area, this research utilized 2022 Beijing-2 remote sensing imagery, 2021 Third National Land Survey data, and second-class forest resource survey data from 2009 and 2019 as data sources. A hierarchical approach was employed to extract land-use classes step by step. Spectral and texture features were extracted from the Beijing-2 imagery, and a random forest classifier was used to identify pure Camellia oleifera plantations. Change information of Camellia oleifera parcels was derived from the second-class forest survey data and combined with visual interpretation to extract mixed Camellia oleifera forests. The classification results were evaluated for accuracy using field survey samples. The results showed an overall classification accuracy of 76% and a Kappa coefficient of 0.63, indicating a high consistency between the classification results and the actual distribution of land features. The area of pure Camellia oleifera plantations in Liaoshi town was 23.79 km², accounting for 30.45% of the total area, with a blocky distribution in the northern and western regions. The area of mixed Camellia oleifera forests was 23.51 km², accounting for 30.10% of the total area, with a patchy distribution in the southeastern and southern regions. This study proposed a novel method for precise identification of Camellia oleifera forests by integrating multi-source remote sensing data. The method shows potential for application in regions with similar growth conditions for Camellia oleifera and provides important reference value for the refined management and scientific planning of Camellia oleifera plantations.