 In recent years, hyperspectral image classification techniques have been gaining popularity among researchers due to their ability to accurately model the development of cities and provide valuable references for urban planning and construction. However, due to the difficulty in acquiring hyperspectral images, only a small number of pixels can be used as training samples. This poses a challenge in terms of extracting and utilizing the spatial and spectral information of these images. To address this issue, we propose a novel hyperspectral image classification method based on dense pyramidal convolution and multi-feature fusion, DPCMF. Dense pyramid convolutions and non-local blocks are used to extract both spatial and spectral features from image samples. These features are then fused together and fed into fully connected layers to produce classification results. Experimental results demonstrate that our method achieves significant improvements over other state-of-the-art methods such as SVM, SSRN, FDSSC, DBMA and DBDA. As a result, our method provides more accurate and reliable terrain and environmental conditions for urban planning, design, construction and management. This article was authored by Jun Sanjiang, Li Xiao, Hong Xiaojiang, and others. We are article.tv, links in the description below.