CN-121482622-B - Enterprise data processing method, device, equipment and medium based on satellite remote sensing

CN121482622BCN 121482622 BCN121482622 BCN 121482622BCN-121482622-B

Abstract

The disclosure provides an enterprise data processing method, device, equipment and medium based on satellite remote sensing, and relates to the technical fields of multi-mode remote sensing driven enterprise-industrial chain intelligent modeling methods and the like. The method comprises the steps of determining factory area outline areas of enterprises in a remote sensing image based on lists of the enterprises, acquiring corresponding time sequence remote sensing data acquired by various types of satellites based on the factory area outline areas, determining pixel numbers of the enterprises based on the factory area outline areas and the time sequence remote sensing data, determining remote sensing feature time sequence values of the enterprises based on the pixel numbers, the factory area outline areas and the time sequence remote sensing data, classifying the enterprises in industry categories, determining remote sensing feature subsets of enterprise subsets in the industry categories based on the remote sensing feature time sequence values, and predicting enterprise data of the enterprise subsets based on the remote sensing feature subsets to obtain enterprise prediction results.

Inventors

DAI YUE
ZHENG YONG
YU LEI
LIU YUTING
WU WENBIN
FAN ZHAOXIN

Assignees

北京四象爱数科技有限公司

Dates

Publication Date: 20260512
Application Date: 20250904

Claims (6)

1. A method for processing enterprise data based on satellite remote sensing, the method comprising: Determining a factory contour area of each enterprise in the remote sensing image based on the lists of the enterprises; acquiring corresponding time sequence remote sensing data acquired by various satellites based on the plant area outline area, wherein the time sequence remote sensing data comprises night lamplight, earth surface daytime temperature, earth surface night temperature, nitrogen dioxide, sulfur dioxide and carbon monoxide raster data; Determining the pixel number of each enterprise based on the factory area outline area and the time sequence remote sensing data comprises the steps of performing spatial superposition on the time sequence remote sensing data and the factory area outline area to obtain a superposition image; Determining remote sensing characteristic time sequence values of all enterprises based on the pixel number, the factory outline area and the time sequence remote sensing data, wherein the determining of the remote sensing characteristic time sequence values of all enterprises comprises dividing the enterprises into small enterprises and large and medium enterprises based on the pixel number, calculating a resonance index based on time sequence observation values extracted from the time sequence remote sensing data, clustering the factory outline area based on the resonance index, determining a primary and secondary activity area, taking a pixel set in a set range with the small enterprises as a center as a buffer area, determining space weights in the buffer area according to the distance from each pixel to the center of the buffer area, calculating to obtain the remote sensing characteristic time sequence values of the small enterprises based on the space weights and the time sequence remote sensing data of the small enterprises, extracting the remote sensing characteristic time sequence values of the large and medium enterprises from the time sequence remote sensing data based on the primary and secondary activity area, taking the remote sensing characteristic time sequence values of the small enterprises and the remote sensing characteristic time sequence values of the large and the medium enterprises as the remote sensing characteristic time sequence values of all enterprises, and processing the time sequence characteristic time sequence values of all enterprises based on the time sequence remote sensing characteristic time sequence observation values extracted from the small enterprises by using the pixel set in the set range as a set range, wherein the pixel values extracted from the small enterprises are extracted by adopting the resonance index, and the time sequence information to obtain the time sequence index values; dividing the industries of the enterprises, and determining remote sensing feature subsets of enterprise subsets in each industry category based on the remote sensing feature time sequence values; And predicting enterprise data of the enterprise subset based on the remote sensing feature subset to obtain an enterprise prediction result, wherein the enterprise prediction result comprises a carbon emission estimated value.
2. The method of claim 1, wherein determining the factory floor contour area for each business in the remote sensing image based on the list of the plurality of businesses comprises: Determining the geographic position and the remote sensing image of each enterprise in the enterprises based on the lists of the enterprises; And obtaining the factory outline area of each enterprise in the remote sensing image based on the geographic position and the remote sensing image.
3. The method of claim 1, wherein predicting enterprise data for the subset of enterprises based on the subset of remote sensing features, the predicting comprising: and inputting the remote sensing feature subset into a pre-trained enterprise data model to obtain an enterprise prediction result output by the enterprise data model.
4. An enterprise data processing apparatus based on satellite remote sensing, the apparatus comprising: The area determining unit is configured to determine the factory outline area of each enterprise in the remote sensing image based on the lists of the plurality of enterprises; The acquisition unit is configured to acquire corresponding time sequence remote sensing data acquired by a plurality of types of satellites based on the factory contour area, wherein the time sequence remote sensing data comprise night light, earth surface daytime temperature, earth surface night temperature, nitrogen dioxide, sulfur dioxide and carbon monoxide grid data; The quantity determining unit is configured to determine the pixel quantity of each enterprise based on the factory area outline area and the time sequence remote sensing data, and is configured to spatially superimpose the time sequence remote sensing data and the factory area outline area to obtain a superimposed image; A numerical value determining unit configured to determine a remote sensing feature timing numerical value of each enterprise based on the number of pixels, the factory area outline area, and the timing remote sensing data; the numerical value determining unit is further configured to divide the plurality of enterprises into small enterprises and large and medium enterprises based on the pixel number, calculate resonance indexes based on time sequence observation values extracted from the time sequence remote sensing data, cluster the outline area of the factory area based on the resonance indexes, determine primary and secondary active areas, take pixel sets in a set range with the small enterprises as centers as buffer areas, determine space weights according to the distances from each pixel to the centers of the buffer areas in the buffer areas, calculate the time sequence numerical values of the remote sensing features of the small enterprises based on the space weights and the time sequence remote sensing data of the small enterprises, extract the time sequence numerical values of the remote sensing features of the large and medium enterprises from the time sequence remote sensing data based on the primary and secondary active areas, take the time sequence numerical values of the remote sensing features of the small enterprises and the time sequence numerical values of the remote sensing features of the large and medium enterprises as the time sequence numerical values of the remote sensing features of each enterprise, and further configured to extract the time sequence observation values from the time sequence remote sensing data by adopting a value extraction grid method, process the time sequence observation values of the pixels under each mode, and perform standard processing on the time sequence numerical values of the pixels to obtain the resonance indexes; the subset determining unit is configured to divide industry categories of the enterprises and determine remote sensing feature subsets of enterprise subsets in each industry category based on the remote sensing feature time sequence numerical values; and the prediction unit is configured to predict enterprise data of the enterprise subset based on the remote sensing feature subset to obtain an enterprise prediction result, wherein the enterprise prediction result comprises a carbon emission estimated value.
5. An electronic device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 3 when the computer program is executed.
6. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 3.

Description

Enterprise data processing method, device, equipment and medium based on satellite remote sensing Technical Field The disclosure belongs to the technical field of data processing, and particularly relates to the technical fields of image processing, deep learning and the like. In particular to an enterprise data processing method and device based on satellite remote sensing, electronic equipment and a computer readable storage medium. Background Currently, dynamic monitoring of enterprise business activities and macro economic indicators has important application value in the fields of government decision, banking wind control, foundation industry, stock market research and judgment and the like. The traditional method mainly adopts structured data sources such as government statistical data, financial reports, questionnaires and the like. The problems of long data acquisition period, low updating frequency, strong hysteresis, coarse spatial resolution and the like generally exist in the methods, and high-frequency dynamic perception of enterprise-level and regional-level economic activities is difficult to realize. In recent years, with the development of satellite remote sensing technology, "alternative data" based on satellite observation data is an emerging information source. The remote sensing data of night light (NTL), surface temperature (LST), air pollutants (such as NO 2、SO2, CO) and the like are proved to have obvious correlation with human activities, energy consumption, industrial emission, economic activity and the like. Therefore, some researches try to estimate urban economy, electricity consumption level or social activity intensity by using a single remote sensing mode (such as noctilucence), but the identification capability of microscopic subjects such as manufacturing enterprises is still limited. In the prior art, attempts have been made to search for macro economic indicators such as GDP, PMI, and industrial increment value using remote sensing data. For example, a regression model is trained using remote sensing images to fit county-level GDP or to make rough estimates of industry yields. However, such methods are limited to urban or industrial macro-scale, lack of spatial localization, dynamic tracking and behavior modeling capabilities at the enterprise level, and cannot realize chain modeling from minimum granularity enterprise production and management activities to macro-economic indexes. In addition, although fusion research on remote sensing multi-source data (such as MODIS, sentinel, VIIRS, TROPOMI and the like) exists, most of the fusion research focuses on the fields of surface coverage classification, climate monitoring, environment assessment and the like, and the fusion research is still weak in aspects of causal mechanism modeling, industry suitability matching, time sequence feature mining and the like between a remote sensing mode and enterprise operation behaviors. Disclosure of Invention The present disclosure provides a data processing method and apparatus, an electronic device, and a computer-readable storage medium. According to a first aspect, an enterprise data processing method based on satellite remote sensing is provided, and the method comprises the steps of determining factory area outline areas of enterprises in a remote sensing image based on lists of the enterprises, acquiring corresponding time sequence remote sensing data acquired by various types of satellites based on the factory area outline areas, determining pixel numbers of the enterprises based on the factory area outline areas and the time sequence remote sensing data, determining remote sensing feature time sequence values of the enterprises based on the pixel numbers, the factory area outline areas and the time sequence remote sensing data, classifying the enterprises in industry categories, determining remote sensing feature subsets of enterprise subsets in the industry categories based on the remote sensing feature time sequence values, and predicting enterprise data of the enterprise subsets based on the remote sensing feature subsets to obtain enterprise prediction results. According to a second aspect, an enterprise data processing device based on satellite remote sensing is provided, and the device comprises an area determining unit, an acquisition unit, a quantity determining unit, a value determining unit and a subset determining unit, wherein the area determining unit is configured to determine factory area outline areas of enterprises in a remote sensing image based on lists of the enterprises, the acquisition unit is configured to acquire corresponding time sequence remote sensing data acquired by multiple types of satellites based on the factory area outline areas, the quantity determining unit is configured to determine pixel quantity of each enterprise based on the factory area outline areas and the time sequence remote sensing data, the value determining unit is configured to