WiMi Hologram Cloud has announced the development of RandomPatchSpatialSpectrumClassifier (RPSSC) technology, which utilises the complementarity between spatial and spectral information in hyperspectral image processing. The technology combines a 2D Gabor filter and a random patch convolution (GRPC) feature extraction method to downscale hyperspectral images and eliminate redundant spectral information, increase inter- and intra-class distance ratios, and prepare data for classification. RPSSC also introduces random patch convolution to extract multilevel spectral features from an image, aiming to synthesise spatial and spectral information. Finally, RPSSC fuses spatial and spectral features to generate a richer feature representation for image classification and employs a support vector machine classifier to achieve accurate classification of hyperspectral images. RPSSC technology has multiple technical advantages, including a simple structure, good adaptability, and the ability to overcome over-smoothing.