 This study proposes a hybrid method for accurately classifying land covers in urban vegetated areas using unmanned aerial vehicle, UAV, remote sensing. The method combines random forest and texture analysis to improve classification accuracy beyond traditional maximum likelihood classifiers. The results show that the inclusion of texture features significantly improves classification accuracy, and that classification accuracy follows an inverted U relationship with texture window size. Overall, this study demonstrates the potential of UAV remote sensing for urban vegetation mapping and highlights the benefits of adopting random forest and texture analysis together to overcome the limitations of off-the-shelf digital cameras. This article was authored by Quan Long Feng, Jin Tao Lu, and Jin Wagong. We are article.tv, links in the description below.