 3D convolutional neural network, 3D, CNN, framework is proposed for accurate hyperspectral image, HSI, classification, which views the HSI cube data altogether without relying on any pre-processing or post-processing, extracting the deep spectral spatial combined features effectively and requires fewer parameters than other deep learning-based methods. The experimental results demonstrate that our 3D, CNN-based method outperform state-of-the-art methods and sets a new record on three real-world HSI data sets captured by different sensors. This article was authored by Eng Lee, Hao Kuizheng, and Chong Chen.