Information Management System Platform for Landscape Engineering
Yuling Zheng, Zhu Wang, Yu Liu · 2024
Traditional landscape engineering management mainly relies on manual recording, which is prone to problems such as delayed information transmission and accumulation of errors, resulting in delayed project progress. The improved Xception model was adopted to automatically identify garden landscapes, and the garden landscape engineering information management system platform was constructed to digitize engineering information. A large number of garden landscape images were collected, denoised, and enhanced. The Xception model was improved, including changing the activation function to the Rectified Linear Unit (ReLU) function and using global maximum pooling to capture features. The results from the test set indicated that the average landscape classification accuracy of the improved Xception model was 98.2%. Building an information management system for landscape engineering through improved Xception can effectively achieve automated landscape classification and improve the accuracy of landscape classification.