Analysis and Research on Color Encoding and K-means Clustering Algorithm in Lingnan Landscape Design

Abstract

As a prominent visual phenomenon, the color environment influences the landscape design of the Lingnan region in three aspects: shape, color, and texture. The purpose of this paper is to examine the landscape nature of color by analyzing the psychological perception characteristics of color and the evolution of Lingnan’s color landscape system. We propose color clustering extraction methods for both single and batch Lingnan landscape images, combining them with color space conversion algorithms like RGB and YIQ to extract colors and segment Lingnan landscape images. We look at the performance parameters of various reconstructed images at various compression ratios. We also draw the performance convergence curves of the clustering algorithm, the JPEG, and the JPEG2000 algorithms to talk about how well they code. Taking the Lingnan region (Guangzhou) as the main object of analysis, we divide the regional color qualities and analyze the evolution and development of landscape color characteristics in the Lingnan region. Most of the Lingnan architectural landscapes have color brightness values in the medium-high range of 5-7. For wall color brightness, most of the brightness class level values are in the high brightness range of 7–10, followed by the medium brightness range of 4–7, and a small portion is in the low brightness range of 1-3.