Research on Application of 3D GIS in Urban Landscape Based on Speech Recognition System

Landscaping and design are designed to meet the needs of people for leisure, entertainment, production and life. The natural landscape is the main part and combined with artificial beautification technology to create a space landscape. Modern planning and design mainly focus on environmental protection and the sustainable development of natural resources. As a natural way of human communication, speech has inherent advantages as a means of human-computer interaction. Automatic speech recognition should enable computers to "understand" human speech and convert speech sequences into text sequences. With the development of deep learning, speech recognition based on deep neural networks has become more and more common. There are currently two frameworks for internal speech recognition modeling: hybrid framework and end-to-end architecture. In this case, this work focuses on the various problems of speech recognition modeling in the two architectures. Based on local topographic maps, remote sensing image maps and various baseline survey data, this paper explores the planning area, uses technology to extract and analyze the spatial data of the research area, and create digital elevation models, remote sensing images, and maps The sounding above forms a virtual three-dimensional environment of the study area. On this basis, the landscape elements are analyzed from the perspectives, topography and spatial pattern to create urban landscapes. The planning and design are based on a comprehensive and intuitive spatial model, which is used to formulate a reasonable anti-urban security plan, with the landscape map of the area as a reference point.