CN-115879594-B - Urban population distribution trend prediction method based on geographic detector

CN115879594BCN 115879594 BCN115879594 BCN 115879594BCN-115879594-B

Abstract

The application provides a city population distribution trend prediction method based on a geographic detector, which comprises the steps of obtaining panoramic street view data of a research area, conducting semantic segmentation to obtain visual perception factors, establishing space perception factors through reachability of different facility types in different travel modes, calculating interpretation rates of the visual perception factors and the space perception factors on population distribution density by using the geographic detector, obtaining a plurality of high interpretation rate perception factors, obtaining corresponding weights based on a judgment matrix, and establishing a population distribution trend index for predicting city population distribution trend. The prediction method of the application takes the perception of the real urban environment as the basis, considers the subjectivity and objectivity of the evaluation, and has higher reference value for urban planning related to urban population distribution.

Inventors

MENG QINGYAN
Qi Junnan
ZHANG LINLIN
HU XINLI

Assignees

中国科学院空天信息创新研究院
海南空天信息研究院

Dates

Publication Date: 20260512
Application Date: 20220906

Claims (7)

1. The urban population distribution trend prediction method based on the geographic detector is characterized by comprising the following steps of: acquiring panoramic street view data of a research area; Performing semantic segmentation on the panoramic street view data to obtain visual perception factors, and establishing at least one urban environment visual perception factor, wherein a neural network model is used for performing semantic segmentation on the panoramic street view data to obtain the visual perception factors, the neural network model is a DeepLabV model trained by utilizing a CITYSCAPE training set, the at least one urban environment visual perception factor is established based on the type and diversity characteristics of the visual perception factors, and the urban environment visual perception factors comprise 7 urban environment visual perception factors including greening, openness, enclosing degree, motorized degree, humanized degree, SIDI diversity and SHDI diversity; Acquiring accessibility of at least one facility type in at least one travel mode based on the panoramic street view data, and establishing at least one urban environment space perception factor, wherein the accessibility of the at least one facility type in the at least one travel mode is acquired by acquiring an isochronous travel range based on the panoramic street view data; Calculating an interpretation rate of the at least one urban environment visual perception factor and the at least one urban environment spatial perception factor on the population distribution density using a geographic detector; Determining at least one high-interpretation-rate perception factor from the at least one urban environment visual perception factor and the at least one urban environment space perception factor based on the interpretation rate, and obtaining a weight corresponding to the high-interpretation-rate perception factor based on a judgment matrix; And establishing a settlement intention index for predicting urban settlement population distribution trend based on the high interpretation-rate perception factors and the corresponding weights.
2. The method of claim 1, wherein the acquiring panoramic street view data of the investigation region comprises: And acquiring panoramic street view data of the research area by setting sampling points with equal intervals or unequal intervals.
3. The prediction method according to claim 2, wherein the obtaining panoramic street view data of the research area by setting equidistant or unequal-distance sampling points includes: And acquiring hundred-degree street view panoramic data of a research area by setting 200m equidistant sampling points of a road network, and cutting and extracting 1/3 of the middle in the vertical direction.
4. The prediction method according to claim 1, wherein the obtaining the isochronous trip range based on the panoramic street view data, obtains reachability of at least one facility type in at least one trip mode, includes: Invoking Isochrone API to obtain travel ranges of each block at 15min and the like under three travel modes of walking, riding and driving; Counting the number of point positions of each type of POI in each equal-time travel range, and calculating reachability of different facilities under different travel modes of the neighborhood based on an accumulated opportunity method; the establishing at least one urban environment space perception factor based on the accessibility of at least one facility type in the at least one travel mode comprises the following steps: based on different travel modes and facility types, the accessibility of office, traffic, business, residence, scientific education, health and green land and square 6-class service facilities under 3 travel modes of walking, riding and driving are calculated respectively, and 18 urban environment space perception factors are obtained.
5. The prediction method according to claim 1, wherein the determining at least one high interpretation-rate perceptual factor from the at least one urban environment visual perceptual factor and the at least one urban environment spatial perceptual factor based on the interpretation-rate, and obtaining the weight corresponding to the high interpretation-rate perceptual factor based on a judgment matrix, comprises: selecting at least one high-interpretation-rate perception factor from the at least one urban environment visual perception factor and the at least one urban environment space perception factor for constructing a hierarchical model; Determining the judgment matrix scale between every two corresponding perception factors according to the difference of the interpretation rate, and obtaining the judgment matrix of the perception factor with high interpretation rate; And obtaining the weight corresponding to the high-interpretation-rate perception factor by resolving the judgment matrix.
6. An electronic device for a geographic detector-based urban population distribution trend prediction method, comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the prediction method according to any one of claims 1 to 5 when executing the program.
7. A medium having stored thereon a computer program for a geographic detector based urban population distribution trend prediction method, which program when executed by a processor implements the prediction method according to any one of claims 1 to 5.

Description

Urban population distribution trend prediction method based on geographic detector Technical Field The invention relates to the technical field of urban remote sensing, in particular to a urban population distribution trend prediction method based on a geographic detector. Background The large population migration activities associated with urbanization are constantly updating the population structure in cities, and large-scale population migration creates tremendous pressure on limited urban resources and environmental capacity. The prediction of fine-scale distribution of the population currently populated remains an important issue that has not yet been addressed. The perception of urban environments profoundly affects the priority of colonisation choices and the stability of occupancy. The public space quality of urban environment and the construction of service facilities positively influence the living wish of people and have close relation with the distribution of living population. Population colonisation is mostly studied based on questionnaire investigation data to analyze living preference of people for urban environment, and quantitative measure study on urban environment perception is incomplete. The development of information types of geospatial data and processing technology greatly enriches the characterization method of the urban environment features under the observation of human visual angles. In recent years, emerging data are widely applied to urban environment characterization research related to human activities, such as POIs (point of information OR point of interest, any non-geographically significant points on a map) are often used for urban morphology characterization, urban functional area identification and gridding population drawing, street view data are widely used for street quality evaluation and urban functional morphology characterization, in addition, a network location service platform is an important supplement to the current accessibility research, and the significance of POIs is that users are connected with geographically significant points under the condition that the knowledge of the users on geographic positions and surrounding information is inaccurate, and the next behavior conversion is carried out. . Population data is mostly dependent on manual investigation, cannot meet spatial continuity and is difficult to collect in a large scale in a short time. The current spatial population distribution forecast subjects are mainly the general population, and the population of colonisation and the population of flow are not considered. Grid cells used in demographic mapping cannot match irregularly shaped neighborhood cells in which people live. Because of the problems of privacy protection or unclear regional unit division boundaries, census data is often in a larger range of general population data. Taking china as an example, the spatial accuracy of urban research related to human activity is mostly limited to the street scale, i.e. the highest accuracy of census. With the improvement of urban environment research precision, the neighborhood is gradually used as a representative of fine research units inside cities in urban landscapes and urban planning analysis. The block is a basic constituent unit of a city structure, which is an important dividing unit related to population activities, as a land unit divided by a road network and having a relatively homogeneous socioeconomic function. In summary, for various and complex urban environments, a set of quantitative analysis-based trend prediction method for urban population distribution with fine scale is needed to more scientifically and objectively guide urban planning and assist urban sustainable development. Disclosure of Invention The application aims to solve the defects existing in the prior art. Aiming at the problem that the quantitative degree of the existing research on urban population space distribution is insufficient, the application aims to provide a urban population distribution trend prediction method based on a geographic detector. On the basis of carrying out quantitative analysis on urban environment perception factors and urban colonisation population distribution, colonisation intention indexes which are both subjective and objective are constructed. According to the urban population distribution trend prediction method based on the geographic detector, the urban population distribution trend prediction method based on the geographic detector comprises the steps of obtaining panoramic street view data of a research area, conducting semantic segmentation on the panoramic street view data to obtain visual perception factors, establishing at least one urban environment visual perception factor, obtaining accessibility of at least one facility type in at least one traveling mode based on the panoramic street view data (accessibility in graph theory refers to the easiness from one vertex to the ot