The Wisdom of Building Dong Wind and Rain Bridges in Northern Guangxi —— An Example from Pingyan Village in Sanjiang County

In this paper, we choose some methods to study the spatial distribution, such as the standard ellipse method, kernel density estimation method, spatial autocorrelation analysis, topographic position index, geographic connection, and other mathematical models to analyze the distribution of Dong wind and rain bridges in northern Guizhou. In order to analyze the construction wisdom of Dong wind and rain bridges, this paper designs a function with wind and rain bridge images as the target and constructs a convolutional neural network-based wind and rain bridge image generation model by encoding and decoding the defined target and adjusting the hyperparameters. After analysis, 21 Dong villages in northern Gui are distributed in the range of 122~749m above sea level, with an average altitude of 385.19m and a kernel density index of 0.3. The model value of this paper can reach up to about 6.2, which can clearly generate images of wind and rain bridges, and then analyze the construction wisdom of wind and rain bridges.