A Quantitative Method of Street Spatial Quality Based on Cross-connection CNN+svr

Abstract

The invention provides a method for quantifying street space quality based on cross-connected CNN+SVR, which mainly involves using cross-connected convolutional neural network to extract street view picture features, and using the extracted features as input features of support vector regression model to quantify street space quality . The method includes: collecting and processing street view pictures to create a corresponding data set, and then training a cross-connected CNN+SVR network to obtain a parameter model of the network, and using the parameter model to quantify the street space quality. The invention gives full play to the advantages of machine learning, reduces the huge workload of researchers in the investigation of street space quality, provides important data support for related research, and provides a new idea for studying street space quality in the field of urban planning.

Description

technical field [0001] The invention relates to the problem of quantitative analysis of street space quality in the field of image analysis, in particular to a method for quantifying street space quality based on machine learning. Background technique [0002] The combination of artificial intelligence with medical care, education, environmental governance, and urban planning will greatly promote the precision of urban public services and comprehensively improve people's quality of life. At present, paying attention to the construction of smart cities is the current hot direction in my country. Urban public space mainly includes multi-functional areas such as streets, shopping malls, squares and parks, and the streets are equivalent to the “skeleton” of the city. As a stage to show the city's economy and life, the street is also an important window to highlight local characteristics. Good street space quality can not only form a friendly and harmonious neighborhood living ...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a street space quality quantification method based on cross-connected CNN+SVR, which combines machine learning with street space quality attributes, fully utilizes the advantages of machine learning self-learning, and can solve the problem of current researchers investigating street space quality. workload problem

Images