Research on Landscape Style Recognition and Design Based on Transfer Learning

Abstract

This paper focuses on the study of landscape style recognition and design based on transfer learning, aiming at the problems such as inaccurate identification, incomplete summary and inaccurate induction in the recognition statistics with photography and naked eyes. Deep neural network and convolutional neural network were used for feature extraction and classification, so as to realize automatic recognition and design of different styles of landscape architecture. Through the training of a large number of data and the deep mechanical shooting technology, the characteristics and rules of different garden styles were analyzed, and the model was built for prediction and classification. And the algorithm is applied to the recognition of garden style, and a new garden recognition system is established. After the experiment, it is found that compared with the original mechanical induction, the new landscape style recognition design system has better universality and accuracy. In terms of style improvement and confirmation, compared with the original recognition success rate, 7.79% was increased. The final results show that the new landscape recognition model based on transfer learning has higher accuracy and application value.