CN-121563015-B - Remote sensing comprehensive monitoring method, device, terminal and medium for natural resources

CN121563015BCN 121563015 BCN121563015 BCN 121563015BCN-121563015-B

Abstract

The application provides a natural resource remote sensing comprehensive monitoring method, a device, a terminal and a medium, wherein the natural resource remote sensing comprehensive monitoring method comprises the steps of obtaining time sequence remote sensing data of a target area, preprocessing, obtaining quantity information of respective natural resources in the target area based on the time sequence remote sensing data to serve as natural resource quantity distribution results, obtaining quality information of the respective natural resources in the target area based on the time sequence remote sensing data to serve as natural resource quality results, obtaining ecological analysis results based on the natural resource quantity distribution results and the natural resource quality results, and outputting the natural resource quantity distribution results, the natural resource quality results and the ecological analysis results to serve as natural resource remote sensing monitoring results. And acquiring interaction relations among the natural resources based on the natural resource quantity distribution results and the natural resource quality results so as to reflect the association among the natural resources, thereby facilitating the analysis of the target area on the whole and realizing more comprehensive monitoring effect.

Inventors

FENG CUNJUN
ZHOU WEI
MA YAN
YU HUI
ZHAN YUANZENG
JIN LIQIANG
WANG XINGKUN
JIN PANPAN
ZHAO JIANXUE
LI JIAXIN
ZHANG YAN
DENG XIAOYUAN

Assignees

浙江省测绘科学技术研究院

Dates

Publication Date: 20260512
Application Date: 20260122

Claims (8)

1. A remote sensing comprehensive monitoring method for natural resources comprises the following steps: Acquiring time sequence remote sensing data of a target area and preprocessing the time sequence remote sensing data; Based on the time sequence remote sensing data, acquiring the quantity information of the respective natural resources in the target area through a natural resource identification model as natural resource quantity distribution results; The method comprises the steps of establishing a remote sensing characteristic index set, calculating and obtaining land remote sensing characteristic index data by combining the remote sensing characteristic index set on the basis of the time sequence remote sensing data in a step-by-step calculation mode, inputting the land remote sensing characteristic index data into a trained regression-physical model to obtain quality information of natural resources in a target area to serve as natural resource quality achievements, wherein the remote sensing characteristic index set comprises a plurality of types of remote sensing characteristic indexes, the regression-physical model comprises a regression sub-model and a physical neural network sub-model, the regression sub-model and the physical neural network sub-model are connected together through a loss function, the regression sub-model is a model based on a CNN framework, and the physical neural network sub-model is a model based on a PINN framework; The method comprises the steps of obtaining initial graph structure data based on natural resource quantity distribution results and natural resource quality results, extracting interaction relations among natural resources based on the initial graph structure data, updating the initial graph structure data based on the interaction relations among the natural resources to obtain updated graph structure data, carrying out feature extraction based on the updated graph structure data to obtain a plurality of global features and local features, obtaining ecological analysis results based on the global features and the local features through a graph structure neural network, wherein the step of extracting the interaction relations among the natural resources based on the initial graph structure data comprises the steps of carrying out time sequence convolution based on the initial graph structure data to obtain a plurality of node change feature vectors, carrying out graph convolution based on the node change feature vectors to obtain edge feature vectors among nodes, and taking the edge feature vectors as the interaction relations; And outputting the natural resource quantity distribution result, the natural resource quality result and the ecological analysis result as natural resource remote sensing monitoring results.
2. The method according to claim 1, wherein the calculating, based on the time-series remote sensing data, in combination with the remote sensing feature index set, by using a step-by-step calculation method, includes: Calculating values of the remote sensing characteristic indexes of various types corresponding to each pixel based on the time sequence remote sensing data, and obtaining pixel remote sensing characteristic index data; Performing pixel segmentation based on the pixel remote sensing characteristic index data to obtain each object remote sensing characteristic and obtain object remote sensing characteristic index data; Clustering the remote sensing features of each object based on the quantity distribution result to obtain the remote sensing features of each land parcel and obtain the index data of the remote sensing features of each land parcel.
3. The method of claim 1, wherein said constructing a remote sensing feature index set comprises: Acquiring each sampling grid based on the quantity information of the natural resources in the natural resource quantity distribution result, and carrying out field sampling based on each sampling grid to acquire actual measurement data corresponding to each sampling grid; Based on the time sequence remote sensing data, combining each sampling grid, calculating the value of each kind of initial remote sensing characteristic index corresponding to each sampling grid; and carrying out association degree analysis on each kind of initial remote sensing characteristic indexes based on the measured data so as to screen each kind of initial remote sensing characteristic indexes meeting association degree requirements as each remote sensing characteristic index, and forming the remote sensing characteristic index set based on each screened remote sensing characteristic index.
4. The method of claim 1, wherein updating the initial graph structure data based on the interaction relationship between the natural resources to obtain updated graph structure data comprises: Acquiring ecological monitoring data; Extracting corresponding ecological information from the ecological monitoring data for each node in the initial graph structure data, and updating the ecological information serving as node characteristics of the corresponding node; based on the interaction relation among the natural resources, acquiring the mutual relation among all nodes in the initial graph structure data, and updating the edge characteristics among the corresponding nodes based on the mutual relation among all nodes.
5. The method of claim 1, wherein after the obtaining the quality information of each of the natural resources in the target area, further comprising: acquiring quality change driving information of the target area, and inputting the quality change driving information and the quality information of each natural resource into a geographic interpretation model; The geographic interpretation model carries out cause analysis on the quality change condition of each natural resource based on the quality change driving information so as to update and output the quality information of each natural resource; and taking the quality information of each natural resource output after updating as the quality achievement of the natural resource.
6. The natural resource remote sensing comprehensive monitoring device is characterized by comprising a preprocessing module, a quantity analysis module, a quality analysis module, an ecological analysis module and an output module; the preprocessing module is used for acquiring time sequence remote sensing data of the target area and preprocessing the time sequence remote sensing data; The quantity analysis module is used for acquiring quantity information of respective natural resources in the target area through a natural resource identification model based on the time sequence remote sensing data, and taking the quantity information as a natural resource quantity distribution result; The quality analysis module is used for constructing a remote sensing characteristic index set, calculating and obtaining the remote sensing characteristic index data of the land by combining the remote sensing characteristic index set and adopting a step-by-step calculation mode, inputting the remote sensing characteristic index data of the land into a trained regression-physical model to obtain the quality information of each natural resource in the target area as a natural resource quality result, wherein the remote sensing characteristic index set comprises a plurality of types of remote sensing characteristic indexes, the regression-physical model comprises a regression sub-model and a physical neural network sub-model, the regression sub-model and the physical neural network sub-model are connected together through a loss function, the regression sub-model is a model based on a CNN frame, and the physical neural network sub-model is a model based on a PINN frame; The ecological analysis module is used for acquiring initial graph structure data based on the natural resource quantity distribution result and the natural resource quality result, extracting interaction relation among the natural resources based on the initial graph structure data, updating the initial graph structure data based on the interaction relation among the natural resources to acquire updated graph structure data, carrying out feature extraction based on the updated graph structure data to acquire a plurality of global features and local features, and acquiring ecological analysis results through a graph structure neural network based on the global features and the local features, wherein the extracting the interaction relation among the natural resources based on the initial graph structure data comprises carrying out time sequence convolution based on the initial graph structure data to acquire a plurality of node change feature vectors, carrying out graph convolution based on the node change feature vectors to acquire edge feature vectors among the nodes, and taking the edge feature vectors as the interaction relation; the output module is used for outputting the natural resource quantity distribution result, the natural resource quality result and the ecological analysis result as natural resource remote sensing monitoring results.
7. The terminal is characterized by comprising a processor and a memory, wherein the memory is in communication connection with the processor; The memory is used for storing a computer program, and the processor is used for executing the computer program stored in the memory, so that the terminal executes the natural resource remote sensing integrated monitoring method as claimed in any one of claims 1 to 5.
8. A computer storage medium storing a computer program, wherein the computer program when executed by a processor implements the natural resource remote sensing integrated monitoring method according to any one of claims 1 to 5.

Description

Remote sensing comprehensive monitoring method, device, terminal and medium for natural resources Technical Field The application belongs to the field of remote sensing monitoring, relates to a natural resource remote sensing monitoring technology, and particularly relates to a natural resource remote sensing comprehensive monitoring method, device, terminal and medium. Background The natural resource monitoring refers to the process of observing, evaluating and analyzing the current status of natural resources such as cultivated land, water area, mineral products, forests, grasslands, wetlands, oceans and the like and the dynamic changes thereof. The remote sensing monitoring technology can acquire the ground surface conditions in a large range, long term and periodically, and compared with the traditional monitoring mode relying on manpower field investigation, the remote sensing monitoring technology is combined for natural resource monitoring, so that the efficient and accurate monitoring effect can be realized. The remote sensing monitoring of natural resources comprises monitoring the conditions of quantity, quality, ecological conditions and the like. Currently, natural resources are monitored in terms of quantity, quality, ecological conditions and the like, and are usually obtained through analysis, identification and acquisition through different models. For example, the trained target detection model is adopted to identify the space distribution area and the position of the natural resources so as to realize quantity monitoring, the quality condition of the natural resources is obtained through inversion or regression models, and the analysis model is adopted to input data of various animals, plants, environments and the like so as to analyze and obtain the ecological condition of the current area range. However, the above process of monitoring natural resources separates the process of acquiring the number, quality and ecological conditions, and cannot acquire the correlation between the number, quality and ecological conditions, so that it is difficult to analyze the situation of the natural resources integrally in terms of number, quality and ecological conditions, for example, the number, quality change of a certain type of natural resources can cause the number, quality and ecological conditions in the whole area of other types of natural resources to change, and the above natural resource monitoring process is difficult to accurately acquire the correlation between the transformations, so that the effect of natural resource monitoring is poor. Meanwhile, when the quality condition of natural resources is acquired through a regression model pair, the regression model generally needs to be trained through a large number of samples to improve the accuracy. However, in actual production, the number of samples may be small, resulting in lower accuracy of the regression model and lower accuracy of the quality of the natural resources obtained. Disclosure of Invention The application aims to provide a remote sensing comprehensive monitoring method, device, terminal and medium for natural resources, which are used for solving the problems that the acquisition modes of quantity, quality and ecological conditions in the prior art are mutually independent, and the correlation between the quantity, quality and ecological conditions cannot be acquired, so that the effect of natural resource monitoring is poor. According to a first aspect, the application provides a natural resource remote sensing comprehensive monitoring method, which comprises the steps of obtaining time sequence remote sensing data of a target area, preprocessing the time sequence remote sensing data, obtaining quantity information of respective natural resources in the target area through a natural resource identification model based on the time sequence remote sensing data, constructing a remote sensing characteristic index set, calculating and obtaining local remote sensing characteristic index data by combining the remote sensing characteristic index set in a step-by-step calculation mode based on the time sequence remote sensing data, inputting the local remote sensing characteristic index data into a trained regression-physical model to obtain quality information of the natural resources in the target area, wherein the remote sensing characteristic index set comprises a plurality of types of remote sensing characteristic indexes, extracting interaction relation between the natural resources based on the initial map structural data, updating the initial map structural data based on the interaction relation between the natural resources, obtaining updated map structural data based on the updated map structural data, performing global characteristic extraction and the local characteristic analysis, and the natural resource quality analysis result is obtained through the local characteristic and the natural resource quality analysis