CN-120808164-B - Multispectral remote sensing disaster monitoring and early warning system utilizing deep learning

CN120808164BCN 120808164 BCN120808164 BCN 120808164BCN-120808164-B

Abstract

The invention discloses a multispectral remote sensing disaster monitoring and early warning system utilizing deep learning, which comprises a spectrum data acquisition and preprocessing module, a pattern recognition module, a disaster risk assessment module, a disaster risk early warning module and a closed loop monitoring module, wherein the spectrum data acquisition and preprocessing module is used for continuously recording image data in different time periods based on earth surface reflection information to obtain an original data set and preprocessing the original data set to obtain a clear image data set, the pattern recognition module is used for carrying out pattern classification on the clear image data set to obtain a potential disaster related characteristic combination, the disaster risk assessment module is used for carrying out time sequence analysis on the potential disaster related characteristic combination to obtain a probability assessment value associated with disaster risk, the disaster risk early warning module is used for constructing a quantitative index system associated with disaster risk based on the probability assessment value and determining an early warning reference standard based on the quantitative index system associated with the constructed disaster risk, and the closed loop monitoring module is used for continuously monitoring multispectral data acquired in real time based on the early warning reference standard to generate early warning information output containing a risk grade.

Inventors

HU BO
NA JING
LI JIA
Chen Xiongle
WU YANG

Assignees

广东工业大学

Dates

Publication Date: 20260505
Application Date: 20250714

Claims (4)

1. A multispectral remote sensing disaster monitoring and early warning system utilizing deep learning is characterized by comprising: The spectrum data acquisition and preprocessing module is used for acquiring earth surface reflection information from a plurality of wave bands, continuously recording image data in different time periods based on the earth surface reflection information to obtain an original data set, and preprocessing the original data set to obtain a clear image data set; the pattern recognition module is used for carrying out pattern classification on the clear image data set to obtain potential disaster related feature combinations; the disaster risk assessment module is used for carrying out time sequence analysis on the potential disaster related characteristic combination to obtain a probability assessment value associated with disaster risk; the disaster risk early warning module is used for constructing a disaster risk associated quantization index system based on the probability evaluation value and determining an early warning reference standard based on the constructed disaster risk associated quantization index system; The closed-loop monitoring module is used for continuously monitoring the multispectral data acquired in real time based on the early-warning reference standard to generate early-warning information output containing risk level; The pattern recognition module comprises a spectrum feature extraction unit and a pattern classification unit, wherein the spectrum feature extraction unit is used for extracting spectrum reflection characteristics of the clear image dataset, constructing a multi-dimensional feature matrix containing band differences, analyzing the change rule of the spectrum reflection differences in different time periods based on the multi-dimensional feature matrix, and obtaining feature distribution results capable of reflecting surface changes; The pattern classification unit comprises a data standardization and preliminary extraction subunit, an abnormal fluctuation separation and marking subunit, a characteristic fusion and clustering subunit and a disaster characteristic identification and output subunit, wherein the data standardization and preliminary extraction subunit is used for carrying out standardization processing on spectrum reflection difference data in characteristic distribution results to obtain a standardized characteristic basis, carrying out preliminary extraction on a spectrum reflection difference pattern by adopting a convolutional neural network model based on the standardized characteristic basis to determine a difference pattern distribution characteristic, the abnormal fluctuation separation and marking subunit is used for separating an abnormal fluctuation part in the difference pattern distribution characteristic by adopting a characteristic screening method to obtain a separated abnormal fluctuation set, judging fluctuation amplitude in the abnormal fluctuation set according to a preset threshold range to obtain a marked fluctuation data set, and carrying out fusion processing on the marked fluctuation data set and environmental background data to judge fused background fluctuation characteristics; The disaster risk assessment module comprises a feature analysis and time sequence analysis unit and a risk probability assessment unit, wherein the feature analysis and time sequence analysis unit is used for carrying out weighting treatment on the potential disaster related feature combination to obtain weighted feature distribution; The expression for calculating the disaster risk probability by the risk probability evaluation unit is as follows: ; Wherein W (delta X) is a dynamic weight matrix, CNN (·) is a trained convolutional neural network, P is disaster risk probability, and X is original multispectral time sequence data.
2. The multi-spectral remote sensing disaster monitoring and early warning system utilizing deep learning according to claim 1, wherein the multi-spectral data acquisition and preprocessing module comprises a multi-spectral data acquisition unit and a data preprocessing unit; the multispectral data acquisition unit is used for acquiring earth surface reflection information from a plurality of wave bands and continuously recording image data in different time periods to form an original data set containing dynamic change characteristics; The data preprocessing unit is used for carrying out denoising and correction operation on each frame of image of the original data set by adopting an image preprocessing technology, and generating an optimized clear image data set.
3. The multi-spectral remote sensing disaster monitoring and early warning system using deep learning according to claim 2, wherein the data preprocessing unit comprises a first preprocessing subunit, a second preprocessing subunit, a third preprocessing subunit, a fourth preprocessing subunit and a fifth preprocessing subunit; The first preprocessing subunit is used for carrying out preliminary denoising processing on each frame of image, eliminating external interference and obtaining a denoised first image data set; the second preprocessing subunit is used for performing geometric and radiation correction on the denoised first image data set to generate a corrected second image data set; The third preprocessing subunit performs secondary denoising processing on the corrected second image data set by adopting a comparison threshold range method to obtain an optimized third image data set; The fourth preprocessing subunit is configured to group the optimized third image dataset according to a time sequence to obtain a grouped fourth image dataset, and extract an image change rule of the grouped fourth image dataset in each time period to obtain a change feature set; And the fifth preprocessing subunit is used for classifying abnormal segments of the change feature set by adopting a random forest algorithm to obtain an optimized clear image data set.
4. The multi-spectrum remote sensing disaster monitoring and early warning system utilizing deep learning according to claim 1, wherein the disaster risk early warning module comprises a quantitative index construction unit, a dynamic early warning threshold setting unit and an early warning information generation and output unit; The quantization index construction unit is used for constructing a quantization index system associated with disaster risks according to the probability evaluation value; the dynamic early warning threshold setting unit is used for setting a dynamically adjusted early warning threshold system according to different types of disaster characteristics and combining application scenes and risk levels; the early warning information generation and output unit is used for analyzing the real-time monitoring data based on the early warning threshold system, triggering an early warning signal when an analysis result accords with an abnormal change mode, and generating and outputting early warning information containing risk level.

Description

Multispectral remote sensing disaster monitoring and early warning system utilizing deep learning Technical Field The invention belongs to the technical field of remote sensing monitoring and disaster early warning, and particularly relates to a multispectral remote sensing disaster monitoring and early warning system utilizing deep learning. Background The multispectral remote sensing technology has a vital role in the disaster monitoring and early warning field, and provides important data support for the prediction and management of natural disasters by capturing spectral information of different wave bands on the ground surface. However, many current disaster monitoring methods have significant shortcomings in data analysis and early warning mechanisms. The traditional method often depends on data at a single time point too, lacks continuous tracking capability for dynamic change, and more importantly, the prior art is poor in performance when processing data interference under a complex environment, for example, factors such as cloud cover, surface coverage change and the like often cause misjudgment or missing report, and the requirement of high-precision early warning is difficult to meet. Against this background, the core challenges facing this field are increasingly manifest. Firstly, the dynamic change characteristic of multispectral data requires that the system can accurately capture the slight difference of the spectral reflectivity in different time periods, and the difference is often interfered by various external factors, so that the recognition difficulty is increased. The more advanced problem is how to relate these differences to the potential risk of disaster occurrence, forming a scientific and reasonable early warning standard. If the problems of difference identification and risk association cannot be effectively solved, the timeliness and the accuracy of the early warning system are greatly reduced, and effective prompt is difficult to provide before the disaster occurs. Therefore, how to accurately identify abnormal change modes in multispectral data through a deep learning technology and establish a scientific and reasonable early warning threshold system based on spectrum difference values becomes a key problem for constructing a high-efficiency disaster early warning system. Disclosure of Invention In order to solve the technical problems, the invention provides a multispectral remote sensing disaster monitoring and early warning system utilizing deep learning, which aims to solve the problems existing in the prior art. In order to achieve the above object, the present invention provides a multispectral remote sensing disaster monitoring and early warning system using deep learning, comprising: The spectrum data acquisition and preprocessing module is used for acquiring earth surface reflection information from a plurality of wave bands, continuously recording image data in different time periods based on the earth surface reflection information to obtain an original data set, and preprocessing the original data set to obtain a clear image data set; the pattern recognition module is used for carrying out pattern classification on the clear image data set to obtain potential disaster related feature combinations; the disaster risk assessment module is used for carrying out time sequence analysis on the potential disaster related characteristic combination to obtain a probability assessment value associated with disaster risk; the disaster risk early warning module is used for constructing a disaster risk associated quantization index system based on the probability evaluation value and determining an early warning reference standard based on the constructed disaster risk associated quantization index system; And the closed-loop monitoring module is used for continuously monitoring the multispectral data acquired in real time based on the early-warning reference standard to generate early-warning information output containing risk level. Optionally, the multispectral data acquisition and preprocessing module comprises a multispectral data acquisition unit and a data preprocessing unit; the multispectral data acquisition unit is used for acquiring earth surface reflection information from a plurality of wave bands and continuously recording image data in different time periods to form an original data set containing dynamic change characteristics; The data preprocessing unit is used for carrying out denoising and correction operation on each frame of image of the original data set by adopting an image preprocessing technology, and generating an optimized clear image data set. Optionally, the data preprocessing unit comprises a first preprocessing subunit, a second preprocessing subunit, a third preprocessing subunit, a fourth preprocessing subunit and a fifth preprocessing subunit; The first preprocessing subunit is used for carrying out preliminary denoising processing on each fra