Search

CN-121997181-A - Mountain area multi-type interference identification method and system based on time sequence segmentation algorithm

CN121997181ACN 121997181 ACN121997181 ACN 121997181ACN-121997181-A

Abstract

The invention provides a mountain area multi-type interference identification method and system based on a time sequence segmentation algorithm, which comprises the steps of obtaining multispectral remote sensing image data covering a target area, carrying out pretreatment and effective pixel screening to form a pretreatment image set, calculating normalized vegetation indexes pixel by pixel and obtaining a continuous vegetation index time sequence data set through time sequence interpolation and reconstruction, constructing an interference sample true value point sequence based on verification sample remote sensing image data, inputting the vegetation index time sequence data set into a Behcet time sequence decomposition algorithm, decomposing the vegetation index time sequence data into trend components, seasonal components and mutation components, extracting the mutation point sequence, calculating a matching rate based on the interference sample true value point sequence and the mutation point sequence, carrying out self-adaption optimization and correction on parameters of the Behcet time sequence decomposition algorithm until the matching rate reaches a preset threshold value, outputting corrected interference detection results, and realizing accurate and automatic identification of mountain area interference information and improving the accuracy and repeatability of interference monitoring.

Inventors

  • Lv Rongfang
  • ZHOU LIANG
  • GAO HONG

Assignees

  • 兰州交通大学

Dates

Publication Date
20260508
Application Date
20260206

Claims (10)

  1. 1. The mountain area multi-type interference identification method based on the time sequence segmentation algorithm is characterized by comprising the following steps of: Acquiring multispectral remote sensing image data covering a target area, preprocessing the multispectral remote sensing image data and screening effective pixels to form a preprocessed image set, wherein the time sequence time of the multispectral remote sensing image data is longer than a preset time length; Calculating normalized vegetation indexes pixel by pixel based on the preprocessed image set, and obtaining a continuous vegetation index time sequence data set through time sequence interpolation and reconstruction; establishing an interference sample truth value point sequence based on verification sample remote sensing image data, wherein the interference sample truth value point sequence comprises interference occurrence time, spatial position and interference type information; inputting the vegetation index time sequence data set into a Behcet time sequence decomposition algorithm, decomposing the vegetation index time sequence data set into a trend component, a season component and a mutation component, and extracting a mutation point sequence as a preliminary interference identification result; And calculating a matching rate based on the interference sample truth value point sequence and the mutation point sequence, performing self-adaptive optimization and correction on parameters of the Bayesian time sequence decomposition algorithm according to the matching rate until the matching rate reaches a preset threshold, and outputting a corrected interference detection result.
  2. 2. The method for identifying multi-type interference in mountain areas based on a time sequence segmentation algorithm according to claim 1, wherein preprocessing and effective pixel screening are performed on the multispectral remote sensing image data to form a preprocessed image set, and the method comprises the following steps: sequentially carrying out radiometric calibration, geometric correction and atmospheric correction on the multispectral remote sensing image to obtain a preprocessed image; Masking noise factors in the preprocessed image by using a masking algorithm to obtain masking results; counting the effective observation quantity of each pixel in all time sequences based on the mask result; Comparing the effective observed quantity with a preset time sequence integrity threshold value; And screening out target pixels with the effective observation quantity being greater than or equal to the time sequence integrity threshold value, and incorporating the effective observation value in the target pixels into the preprocessing image set.
  3. 3. The method for identifying the multi-type interference in the mountain area based on the time sequence segmentation algorithm according to claim 1, wherein the steps of calculating the normalized vegetation index pixel by pixel and obtaining the continuous vegetation index time sequence data set through time sequence interpolation and reconstruction include: calculating a normalized vegetation index pixel by pixel based on the preprocessed image set; And carrying out time sequence interpolation and reconstruction on the normalized vegetation index by adopting a moving median filtering method to obtain the vegetation index time sequence data set.
  4. 4. The mountain area multi-type interference identification method based on a time sequence segmentation algorithm as set forth in claim 1, wherein the constructing an interference sample true value point sequence based on verification sample remote sensing image data comprises: Based on the verification sample remote sensing image data, identifying a region with a change value of the target object larger than a preset change value in a time sequence front-back comparison mode, and extracting an interference event; Labeling the occurrence time, the spatial position and the interference type for each interference event to form a sample truth value point sequence; And merging the points with multiple interferences at the same position in the sample truth point sequence to construct an interference sample truth point sequence with unique spatial position and multi-attribute information.
  5. 5. The mountain area multi-type interference recognition method based on the time sequence segmentation algorithm as claimed in claim 1, wherein the extracting the mutation point sequence as the interference preliminary recognition result comprises: Extracting nested mutation points in the trend component and the season component to form a candidate mutation point set; and performing primary screening on the candidate mutation point set by combining mutation intensity, occurrence probability and trusted interval corresponding to the nested mutation points to obtain a mutation point sequence as the interference primary identification result.
  6. 6. The mountain area multi-type interference recognition method based on the time sequence division algorithm according to claim 1, wherein calculating a matching rate based on the interference sample true value point sequence and the mutation point sequence comprises: calculating time deviation between each pair of detection mutation points and sample truth points; If the time deviation is smaller than or equal to the preset time tolerance, judging that the time deviation is a matching point; and counting the number of all the matching points and the number of the sample true value points, and calculating the matching rate.
  7. 7. The mountain area multi-type interference identification method based on the time sequence segmentation algorithm according to claim 1, wherein the self-adaptive optimization and correction are performed on parameters of the bayesian time sequence decomposition algorithm according to the matching rate until the matching rate reaches a preset threshold, and the method comprises the following steps: When the matching rate is lower than a preset threshold, adjusting key parameters of a Bayesian time sequence decomposition algorithm, wherein the key parameters comprise at least one of a seasonal mutation point number range, a trend mutation point number range, a seasonal segmentation minimum length, a trend segmentation minimum length and a mutation intensity threshold; and repeating the Bayesian time sequence decomposition and the matching rate calculation after updating the key parameters until the matching rate reaches a preset threshold.
  8. 8. The mountain area multi-type interference identification method based on the time sequence division algorithm as claimed in claim 7, wherein adjusting key parameters of the bayesian time sequence decomposition algorithm comprises: adopting a hierarchical self-adaptive parameter tuning mechanism to adjust key parameters of a Bayesian timing decomposition algorithm; The hierarchical self-adaptive parameter tuning mechanism comprises: selecting a representative sample from a target area to perform parameter sensitivity analysis, and determining an initial parameter range; based on the initial parameter range, carrying out regional optimization according to the land type or ecological type difference; and performing local fine adjustment on pixels with consistency check errors greater than preset errors after the regional adjustment.
  9. 9. The method for identifying multi-type interference in mountain areas based on time sequence segmentation algorithm as set forth in claim 8, wherein selecting a representative sample in a target area for parameter sensitivity analysis, determining an initial parameter range, comprises: Selecting a sample point set with space representativeness from the interference sample truth value point sequence by adopting a layered random sampling method according to at least one factor of the type of land, the altitude gradient, the vegetation coverage and the historical interference frequency in the target area; Based on vegetation index time sequence data of the sample point set, performing sensitivity analysis on key parameters of a Behcet time sequence decomposition algorithm by adopting a grid search method and a cross verification method, and determining an initial parameter range; Wherein the process of sensitivity analysis comprises: Sequentially changing the values of the key parameters, recording the change of the matching rate of the sample point set after each change, and calculating the sensitivity of the key parameters to the influence of the matching rate; And screening out key parameters with sensitivity larger than preset sensitivity as parameters to be tuned, and determining the initial parameter range of each parameter to be tuned by combining the physical meaning and the experience range of the parameters.
  10. 10. A mountain area multi-type interference recognition system based on a time sequence segmentation algorithm, which is characterized by comprising: The device comprises a preprocessing module, a target area acquisition module and a target area acquisition module, wherein the preprocessing module is used for acquiring multispectral remote sensing image data covering the target area, preprocessing the multispectral remote sensing image data and screening effective pixels to form a preprocessed image set, and the time sequence time length of the multispectral remote sensing image data is longer than a preset time length; The reconstruction module is used for calculating normalized vegetation indexes pixel by pixel based on the preprocessing image set and obtaining a continuous vegetation index time sequence data set through time sequence interpolation and reconstruction; the construction module is used for constructing an interference sample truth value point sequence based on the verification sample remote sensing image data, wherein the interference sample truth value point sequence comprises interference occurrence time, spatial position and interference type information; The identification module is used for inputting the vegetation index time sequence data set into a Bayesian time sequence decomposition algorithm, decomposing the vegetation index time sequence data set into a trend component, a season component and a mutation component, and extracting a mutation point sequence as a disturbance primary identification result; and the self-adaptive optimization module is used for calculating the matching rate based on the interference sample true value point sequence and the mutation point sequence, carrying out self-adaptive optimization and correction on parameters of the Behcet time sequence decomposition algorithm according to the matching rate until the matching rate reaches a preset threshold value, and outputting a corrected interference detection result.

Description

Mountain area multi-type interference identification method and system based on time sequence segmentation algorithm Technical Field The invention relates to the technical field of geographic information, in particular to a mountain area multi-type interference identification method and system based on a time sequence segmentation algorithm. Background The mountain area interference refers to external effects which are caused by global climate change and human activities under complex mountain terrain and have significant influence on the structure, ecological process and function of the ecological system, such as landslide, debris flow, plant diseases and insect pests, road construction, agriculture and animal husbandry activities and the like. The time, range and type of the mountain area interference are accurately identified, the response process and recovery potential of the mountain area ecological system to external disturbance can be analyzed, and scientific support is provided for mountain area ecological restoration and disaster prevention and control. The traditional interference data depends on modes such as disaster files, ground investigation and the like, has the problems of lag in data update and limited space coverage, and is difficult to form a long-time-sequence continuous monitoring system. Landsat, sentinel and other multi-source satellites provide remote sensing observation data with wide coverage and long time sequence, can reflect vegetation change and interference mutation on the earth surface of a mountain area, but factors such as cloud, fog and shadow in the mountain area easily cause noise and loss of the remote sensing data, and reduce data reliability. The existing time sequence segmentation and mutation detection algorithm can realize partial interference extraction, but lacks a standardized flow in mountain area application, lacks complete specifications from data preprocessing to mutation extraction, lacks a consistency verification mechanism of mutation detection results, relies on experience setting of key parameters of the algorithm, and is difficult to meet high-precision and engineering requirements of mountain area interference identification. Disclosure of Invention The invention provides a mountain area multi-type interference identification method and system based on a time sequence segmentation algorithm, which are used for realizing accurate and automatic identification of the occurrence time, range and type of mountain area interference and improving the accuracy and repeatability of interference monitoring. In one aspect, the invention provides a mountain area multi-type interference identification method based on a time sequence segmentation algorithm, which comprises the following steps: Acquiring multispectral remote sensing image data covering a target area, preprocessing the multispectral remote sensing image data and screening effective pixels to form a preprocessed image set, wherein the time sequence time of the multispectral remote sensing image data is longer than a preset time length; Calculating normalized vegetation indexes pixel by pixel based on the preprocessed image set, and obtaining a continuous vegetation index time sequence data set through time sequence interpolation and reconstruction; establishing an interference sample truth value point sequence based on verification sample remote sensing image data, wherein the interference sample truth value point sequence comprises interference occurrence time, spatial position and interference type information; inputting the vegetation index time sequence data set into a Behcet time sequence decomposition algorithm, decomposing the vegetation index time sequence data set into a trend component, a season component and a mutation component, and extracting a mutation point sequence as a preliminary interference identification result; And calculating a matching rate based on the interference sample truth value point sequence and the mutation point sequence, performing self-adaptive optimization and correction on parameters of the Bayesian time sequence decomposition algorithm according to the matching rate until the matching rate reaches a preset threshold, and outputting a corrected interference detection result. On the other hand, the invention also provides a mountain area multi-type interference identification system based on a time sequence segmentation algorithm, which comprises the following steps: The device comprises a preprocessing module, a target area acquisition module and a target area acquisition module, wherein the preprocessing module is used for acquiring multispectral remote sensing image data covering the target area, preprocessing the multispectral remote sensing image data and screening effective pixels to form a preprocessed image set, and the time sequence time length of the multispectral remote sensing image data is longer than a preset time length; The reconstruction module is