Modern garden design thinking and practice based on spatial information technology

The construction of a garden image classification model is done by combining spatial information technology and analyzing the process of hyperspectral classification of garden images is the main focus of this paper. The linear transformation of the image is performed by principal component analysis to achieve the effect of reducing image dimension. The SVM classifier is used to classify the garden images, and the hyperplane is found in the sample space to distinguish between positive and negative cases. Using a simple linear iterative algorithm, the image superpixels are segmented, and the information contained in them is fused with the features of the hyperspectral image. The positioning of the hyperpixel block impacts the calculation of the mean hyperspectral feature value for each hyperpixel region. The results show that a well-rounded designer needs to achieve 70% aesthetics and 80% rationality to present modern garden design.