Neural Network Model of Urban Landscape Design Based on Multi-Target Detection

Abstract

Urban landscape design is of great significance to the development of the city. To solve the problem of manual acquisition characteristics of urban landscape design systems and low target detection accuracy, a new idea based on neural network and multi-objective testing technology construction urban landscape design system is proposed. By analyzing key technologies, the database establishment proposes to build a city landscape design system. Establish a city landscape design 3D model library using neural network and multi-objective testing technology. The system enables terrain measurements, GIS, visualization, massive data processing, virtual reality technology, etc., so that users can more and more effectively feel the rationality of space design and the feasibility of planning program. Through the experimental test, the following conclusions were obtained. First, the accuracy of multi-objective detection technology is maintained at around 88%. Second, the system landscape generation module generates fast, and the calculation time is between 0.57 and 46 s. Third, through analysis of the evaluation results of ecological suitability, the A-city landscape ecological function partition is divided into the plan for the plan to provide reliable data protection. The fourth is based on satisfaction evaluation indicators, which is conducive to the choice of the optimal plan of urban landscape design, thereby promoting the sustainable development of the city.