Search

CN-121838024-B - Slope video monitoring disaster early warning method and system based on airborne images

CN121838024BCN 121838024 BCN121838024 BCN 121838024BCN-121838024-B

Abstract

The invention provides a method and a system for early warning a slope video monitoring disaster based on an airborne image, which belong to the field of image processing, wherein the method comprises the steps of acquiring a plurality of time sequence gray level images and dividing each image into a plurality of parts; and judging whether early warning is required to be carried out on the later time sequence of each time sequence according to the distribution of the offset angles and the offset amounts of all the parts which have abnormal movement and have no abnormal movement in each gray level image of each area. The invention aims to solve the problem that when disaster early warning is carried out on a side slope through the on-board image of the side slope, the early warning result is inaccurate because natural characteristic points in the image are easily interfered by environmental factors such as illumination, artificial activities and the like.

Inventors

  • LUO TAO
  • ZHOU YONGSHENG
  • GUO QIYOU
  • YU DAYONG
  • Guo Guanmiao
  • WANG PENG
  • LEI JIE
  • CHEN BIN
  • WU ZHENLIN
  • Huo Mingxuan
  • Cai Qinge
  • WEI XUEYONG
  • Shen Shouxiang
  • ZHU DONGCHUN
  • ZHU ZICHENG

Assignees

  • 西安中交公路岩土工程有限责任公司

Dates

Publication Date
20260508
Application Date
20260313

Claims (8)

  1. 1. The side slope video monitoring disaster early warning method based on the airborne images is characterized by comprising the following steps of: acquiring gray images of each area in a side slope to be early-warned on a plurality of time sequences; Dividing each gray level image into a plurality of parts, and acquiring the distance similar part of each part in the gray level image on the previous time sequence from each region in the side slope for early warning in the gray level image on each time sequence according to the position of each part in the gray level image, so as to obtain the similarity of each part in each gray level image and the distance similar part thereof; according to the distribution of the offset angles and the offset amounts of all the parts generating abnormal movement in each gray level image of each area in the side slope for early warning on each time sequence, and the distribution of the positions, the offset angles and the offset amounts of all the parts not generating abnormal movement in the gray level image in the image, the local turbulence degree and the overall consistency of the gray level image of each area in the side slope for early warning on each time sequence, the position of the part generating abnormal movement in the image is determined by the position of the part generating abnormal movement in the gray level image, the distribution of the offset angles and the overall consistency of the parts generating abnormal movement in the gray level image is determined by the position of the part generating abnormal movement in the gray level image; Judging whether each region in the side slope for early warning should be early-warned at the later time sequence of each time sequence according to the local disorder degree and the overall consistency of the gray scale image of each region in the side slope for early warning at each time sequence; the specific steps of obtaining the most similar part, the credibility of the offset, the offset angle and the offset of each part in each gray level image are as follows: the first side slope to be early-warned The first area is at The first time sequence of gray scale image The distance similar part corresponding to the maximum value in the similarity of each part and all the distance similar parts is recorded as the first part in the side slope for early warning The first area is at The first time sequence of gray scale image The most similar part of the individual parts; first in side slope for early warning The first area is at The first time sequence of gray scale image The coordinates of the central pixel point of each part in the gray level image are used as starting points to perform early warning on the first side slope The first area is at The first time sequence of gray scale image The coordinates of the center pixel point of the most similar part of each part in the gray level image are taken as the end points to generate a vector ; Vector The size of (2) is recorded as the first in the side slope for early warning The first area is at The first time sequence of gray scale image Offset of individual parts, vector The included angle between the direction of the car and the horizontal axis to the right is recorded as the first angle in the side slope for early warning The first area is at The first time sequence of gray scale image Offset angles of the individual portions; Acquiring the first side slope for early warning The first area is at The first time sequence of gray scale image The specific calculation formula of the credibility of the offset of each part is as follows: in the formula, Indicating the first side slope to be warned The first area is at The first time sequence of gray scale image The trustworthiness of the offset of the individual parts, Indicating the first side slope to be warned The first area is at The first time sequence of gray scale image The maximum value of the similarity of the individual parts to all of their distance-like parts, Representing the similarity of a reference part to all its distance-like parts divided by The variance of other similarity outside, wherein the reference part is the first side slope for early warning The first area is at The first time sequence of gray scale image A plurality of sections; The specific steps of obtaining the local disturbance degree and the overall consistency of each region in the side slope for early warning in the gray level image on each time sequence are as follows: Presetting an angle threshold and an offset threshold; When early warning is carried out on the first side slope The first area is at The first time sequence of gray scale image Part and the first part for generating abnormal movement The absolute value of the difference value of the offset angles of the parts generating abnormal movement is smaller than or equal to a preset angle threshold value, so that the first side slope to be early-warned is The first area is at The first time sequence of gray scale image The part generating abnormal movement is marked as the first part in the side slope for early warning The first area is at The first time sequence of gray scale image Offset angle similar parts of the parts generating abnormal movement; When early warning is carried out on the first side slope The first area is at The first time sequence of gray scale image Part and the first part for generating abnormal movement The absolute value of the difference of the offset of the parts generating abnormal movement is smaller than or equal to the preset offset threshold value, so that the first side slope to be early-warned is The first area is at The first time sequence of gray scale image The part generating abnormal movement is marked as the first part in the side slope for early warning The first area is at The first time sequence of gray scale image Offset-similar portions of the portions that generate abnormal movement; According to the first side slope with early warning The first area is at The number of offset similar parts and the number of offset angle similar parts of each abnormal movement generating part in the gray level images at the time sequence are equal to each other, so that the first part in the side slope for early warning is obtained The first area is at Local turbulence within the gray scale image at each time sequence; and obtaining the overall consistency of each region in the side slope for early warning in the gray level image at each time sequence according to the positions, the offset angles and the offset distribution of all the parts which do not generate abnormal movement in the gray level image in the image.
  2. 2. The method for warning against a disaster in a slope video surveillance based on-board imaging according to claim 1, wherein the specific step of acquiring a distance-similar portion of each part in a gray scale image at a previous time sequence from each region in the slope to be warned in each time sequence is as follows: on the first side slope for early warning The first area is at Time sequence and the first Matching and aligning gray images on each time sequence; Acquiring the first side slope for early warning The first area is at The first time sequence of gray scale image The part is at the first Corresponding portions within the grayscale image at each time sequence; the first side slope to be early-warned The first area is at The first time sequence of gray scale image The part is at the first The frame of the corresponding part in the gray level image on each time sequence is used as the frame of the initial part; taking the maximum moving distance traversing to the right and downward as the first side slope for early warning by taking the width of the step length as 1 pixel point The first area is at Maximum width of each part in gray level image at time sequence, first in side slope for early warning The first area is at The frames of the initial part are traversed from left to right in the gray level images on the time sequence, the frames of the initial part are traversed from top to bottom, and the part surrounded by the frames of the initial part in the traversing process and the first part in the side slope for early warning are traversed The first area is at The first time sequence of gray scale image The part is at the first The corresponding parts in the gray level images at the time sequence are all marked as the first part in the side slope for early warning The first area is at The first time sequence of gray scale image The distance of the individual parts is similar.
  3. 3. The method for warning the disaster in the slope video surveillance based on the airborne images according to claim 1, wherein the specific steps of obtaining the similarity between each part in each gray level image and the similar part of the gray level image are as follows: Acquiring the first side slope for early warning The first area is at The first time sequence of gray scale image Each part being similar to each of the distances A value; the first side slope to be early-warned The first area is at The first time sequence of gray scale image The part and the first From similar parts The value is recorded as the first in the side slope for early warning The first area is at The first time sequence of gray scale image The part and the first Similarity of the distance-similar parts.
  4. 4. The method for warning the slope video monitoring disaster based on the airborne images according to claim 1, wherein the specific steps of obtaining the portion which does not generate abnormal movement in each gray level image are as follows: constructing a two-dimensional coordinate system with the abscissa being the credibility of deviation and the ordinate being the similarity with the most similar part, and carrying out early warning on the first side slope The first area is at Mapping all parts in the gray level images on each time sequence into a two-dimensional coordinate system to obtain a plurality of data points in the two-dimensional coordinate system; dividing all data points in a two-dimensional coordinate system into outliers and non-outliers by using an LOF algorithm; the corresponding parts of the outliers in all the data points in the two-dimensional coordinate system are marked as the parts for generating abnormal movement; The portion corresponding to the non-outlier point among all the data points in the two-dimensional coordinate system is recorded as the portion where the abnormal movement is not generated.
  5. 5. The method for warning of a slope video surveillance disaster based on-board images according to claim 1, wherein the obtained warning is the first in the slope The first area is at The specific calculation formula of the local disturbance degree in each time sequence gray level image is as follows: in the formula, Indicating the first side slope to be warned The first area is at Local turbulence within the gray scale image at each time sequence, Indicating the first side slope to be warned The first area is at The number of parts generating abnormal movement within the gray-scale image at the timing, Indicating the first side slope to be warned The first area is at The first time sequence of gray scale image The number of offset-like portions of the portion that produces the abnormal movement, Indicating the first side slope to be warned The first area is at The first time sequence of gray scale image The number of offset angle-like portions of the portions that produce the abnormal movement; Is an exponential function with a base of natural constant.
  6. 6. The method for warning the slope video monitoring disaster based on the airborne images according to claim 1, wherein the specific steps of obtaining the overall consistency of each region in the slope for warning in each time sequence in the gray level image are as follows: the first side slope to be early-warned The first area is at Time-series gray scale image inner first Center pixel and the first pixel of the portion without abnormal movement The Euclidean distance of the central pixel point of the part which does not generate abnormal movement is recorded as the first side slope for early warning The first area is at Time-series gray scale image inner first Parts and the first part not generating abnormal movement Euclidean distance of the parts that do not generate abnormal movement; the first side slope to be early-warned The first area is at All parts of the gray level image on the time sequence, which do not generate abnormal movement, are connected with the first part in the side slope for early warning The first area is at Time-series gray scale image inner first The parts which do not generate abnormal movement and have the smallest absolute value of the difference value of the offset angles are marked as the first part in the side slope for early warning The first area is at Time-series gray scale image inner first Similar parts of the parts that do not generate abnormal movement; Acquiring the first side slope for early warning The first area is at The specific calculation formula of the overall consistency in the gray scale images at each time sequence is as follows: in the formula, Indicating the first side slope to be warned The first area is at Overall consistency within the gray scale image over time, Indicating the first side slope to be warned The first area is at The number of parts in the gray-scale image at the timing where no abnormal movement is generated, Indicating the first side slope to be warned The first area is at Time-series gray scale image inner first The offset angle of the portion where the abnormal movement is not generated, Indicating the first side slope to be warned The first area is at Time-series gray scale image inner first The offset angle of the similar portion of the portion where the abnormal movement is not generated, Indicating the first side slope to be warned The first area is at Time-series gray scale image inner first The minimum value in euclidean distance of a portion that does not produce abnormal movement from all of its similar portions, Indicating the first side slope to be warned The first area is at Time-series gray scale image inner first The minimum value in euclidean distance of a portion that does not produce abnormal movement from all of its similar portions, Indicating the first side slope to be warned The first area is at Time-series gray scale image inner first The offset of the portion where the abnormal movement is not generated, Indicating the first side slope to be warned The first area is at Time-series gray scale image inner first The offset of the portion where the abnormal movement is not generated, Representing the function of the absolute value, Is an exponential function with a base of natural constant.
  7. 7. The method for warning the slope video monitoring disaster based on the airborne images according to claim 1, wherein the specific steps for judging whether each area in the slope to be warned should be warned at the later time sequence of each time sequence are as follows: Acquiring the first side slope for early warning The first area is at The specific calculation formula of the degree of expression that produces the abnormal movement tendency at each time sequence is as follows: in the formula, Indicating the first side slope to be warned The first area is at The degree of appearance of abnormal movement tendency is generated in time sequence, Indicating the first side slope to be warned The first area is at Local turbulence within the gray scale image at each time sequence, Indicating the first side slope to be warned The first area is at Overall consistency within the gray scale image over time; presetting a performance threshold value if If the performance threshold value is larger than the performance threshold value, the first side slope is considered to be early-warned The first area is at And (5) generating a different time sequence, and then carrying out early warning.
  8. 8. The system for warning the slope video monitoring disaster based on the airborne images comprises a memory, a processor and a computer program stored in the memory and running on the processor, wherein the steps of the slope video monitoring disaster warning method based on the airborne images as claimed in any one of claims 1 to 7 are realized when the processor executes the computer program.

Description

Slope video monitoring disaster early warning method and system based on airborne images Technical Field The invention belongs to the field of image processing, and particularly relates to a slope video monitoring disaster early warning method and system based on an onboard image. Background The side slope video monitoring disaster early warning based on the airborne images refers to continuous or periodic video/image data acquisition of the side slope by utilizing visible light, infrared or multispectral cameras carried on a flight platform such as an unmanned plane and the like, and the unstable signs of the side slope are identified and monitored in real time or on time through an intelligent analysis technology. The current slope disaster early warning mainly depends on displacement change of natural characteristic points in a monitoring scene, but the natural characteristic points are easily interfered by environmental factors such as illumination, vegetation growth, artificial activities and the like, so that the stability of monitoring data is poor and the false alarm rate is high. The conventional method is difficult to distinguish normal image changes interfered by normal environmental factors from local abnormal displacement of disaster precursors, so that the accuracy of early warning results is insufficient. Disclosure of Invention In order to solve the problem that the current monitoring method is difficult to distinguish normal image changes interfered by normal environmental factors from local abnormal displacement before disasters, so that the accuracy of an obtained early warning result is insufficient, the invention provides a slope video monitoring disaster early warning method and system based on an airborne image. In order to achieve the above object, the present invention provides the following technical solutions: acquiring gray images of each area in a side slope to be early-warned on a plurality of time sequences; Dividing each gray level image into a plurality of parts, and acquiring the distance similar part of each part in the gray level image on the previous time sequence from each region in the side slope for early warning in the gray level image on each time sequence according to the position of each part in the gray level image, so as to obtain the similarity of each part in each gray level image and the distance similar part thereof; according to the distribution of the offset angles and the offset amounts of all the parts generating abnormal movement in each gray level image of each area in the side slope for early warning on each time sequence, and the distribution of the positions, the offset angles and the offset amounts of all the parts not generating abnormal movement in the gray level image in the image, the local turbulence degree and the overall consistency of the gray level image of each area in the side slope for early warning on each time sequence, the position of the part generating abnormal movement in the image is determined by the position of the part generating abnormal movement in the gray level image, the distribution of the offset angles and the overall consistency of the parts generating abnormal movement in the gray level image is determined by the position of the part generating abnormal movement in the gray level image; and judging whether each region in the side slope for early warning should be early-warned at the later time sequence of each time sequence according to the local turbulence degree and the overall consistency of each region in the side slope for early warning at each time sequence. Further, the specific steps of acquiring the distance similar part of each part in the gray level image at the previous time sequence from each region in the side slope for early warning in the gray level image at each time sequence are as follows: on the first side slope for early warning The first area is atTime sequence and the firstMatching and aligning gray images on each time sequence; Acquiring the first side slope for early warning The first area is atThe first time sequence of gray scale imageThe part is at the firstCorresponding portions within the grayscale image at each time sequence; the first side slope to be early-warned The first area is atThe first time sequence of gray scale imageThe part is at the firstThe frame of the corresponding part in the gray level image on each time sequence is used as the frame of the initial part; taking the maximum moving distance traversing to the right and downward as the first side slope for early warning by taking the width of the step length as 1 pixel point The first area is atMaximum width of each part in gray level image at time sequence, first in side slope for early warningThe first area is atThe frames of the initial part are traversed from left to right in the gray level images on the time sequence, the frames of the initial part are traversed from top to bottom, and the part surrounded by the frames