CN-122023825-A - Ice and snow interface characteristic identification and extraction method based on optical image

CN122023825ACN 122023825 ACN122023825 ACN 122023825ACN-122023825-A

Abstract

The invention relates to the technical field of image processing, in particular to an ice and snow interface characteristic identification and extraction method based on optical images. The method comprises the steps of obtaining a corresponding ice and snow area optical image sequence through an unmanned aerial vehicle-mounted spectral camera, synchronously recording shooting positions, attitude angles and ice and snow elevation data, generating an ice and snow area optical image data set corresponding to a space-time tag, carrying out radiation calibration and atmospheric correction on the ice and snow area optical image data set, carrying out image geometric registration simultaneously to obtain an ice and snow area geometric registration image set sequence, carrying out ice and snow area type classification and ice and snow interface post-treatment according to the ice and snow area geometric registration image set sequence to obtain an ice and snow interface contour characteristic curve, and carrying out ice and snow interface dynamic characteristic recognition based on the ice and snow interface contour characteristic curve to generate an ice and snow interface dynamic characteristic monitoring report. The invention can realize high-precision identification of the ice and snow interface characteristics and can improve the monitoring efficiency of ice and snow areas.

Inventors

SUN PENG
TIAN BO
ZHANG ZHICHENG
ZHANG HAN
Zeng Yinggen
CHENG SHUAI
MAO XIAOHONG
LI SILI
LIU FUTAO

Assignees

清华大学苏州汽车研究院（吴江）
交通运输部公路科学研究所
中国极地研究中心(中国极地研究所)

Dates

Publication Date: 20260512
Application Date: 20260129

Claims (10)

1. The ice and snow interface characteristic identification and extraction method based on the optical image is characterized by comprising the following steps of: Step S1, acquiring an optical image sequence of an ice and snow area corresponding to visible light, near infrared and short wave infrared bands through an unmanned aerial vehicle-mounted spectrum camera, synchronously recording shooting positions, attitude angles and ice and snow elevation data, and generating an optical image dataset of the ice and snow area corresponding to a space-time tag based on the shooting positions, the attitude angles and the ice and snow elevation data; S2, performing radiometric calibration and atmospheric correction on the optical image data set of the ice and snow area to generate a radiation correction image data set of the ice and snow area; Step S3, classifying the ice and snow domain types according to the geometric registration image set sequence of the ice and snow region to generate an ice and snow domain type classification map; And S4, carrying out ice and snow interface dynamic feature recognition on the geometric registration image set sequence of the ice and snow area based on the ice and snow interface contour feature curve so as to generate an ice and snow interface dynamic feature monitoring report, wherein the ice and snow interface dynamic feature monitoring report comprises corresponding ablation boundary migration, snow cover change depth and ice crack development degree at the ice and snow interface.
2. The method for identifying and extracting the ice and snow interface features based on the optical image as claimed in claim 1, wherein the step S1 comprises the following steps: s11, obtaining topographic relief data of the ice and snow area pairs, and planning and generating flight scanning tracks of the ice and snow areas corresponding to the unmanned aerial vehicle based on the topographic relief data; S12, carrying a spectrum camera on the unmanned aerial vehicle to select visible light red, green and blue wave bands sensitive to the albedo of the ice and snow, a near infrared wave band capable of reflecting the particle size of the ice and snow and a short wave infrared wave band capable of distinguishing ice and snow from water, and shooting by utilizing the spectrum camera based on the flight scanning track of the ice and snow area corresponding to the unmanned aerial vehicle to obtain an optical image sequence of the ice and snow area corresponding to the visible light, the near infrared and the short wave infrared wave bands; S13, synchronously recording shooting positions, attitude angles and ice and snow elevation data corresponding to shooting moments through integrating an inertial measurement unit and a global navigation satellite system in the unmanned aerial vehicle; And S14, taking the shooting position, the attitude angle and the ice and snow elevation data corresponding to the shooting time as the space-time tag of the optical image of the ice and snow area corresponding to the same shooting time in the optical image sequence of the ice and snow area, so as to generate an optical image data set of the ice and snow area corresponding to the space-time tag.
3. The method for identifying and extracting the ice and snow interface features based on the optical image as claimed in claim 1, wherein the step S2 comprises the following steps: s21, acquiring a camera response function calibrated in a laboratory through a spectrum camera, and performing optical image radiometric calibration on an optical image dataset in an ice and snow area based on the camera response function so as to convert a pixel gray value corresponding to an optical image in the ice and snow area into a gas-meter radiation brightness, so as to obtain the gas-meter radiation image dataset in the ice and snow area; S22, acquiring the atmospheric visibility and the water vapor content corresponding to real-time measurement, and performing atmospheric radiation calculation on the ice and snow area based on the atmospheric visibility and the water vapor content to obtain the atmospheric uplink radiation amount and the atmospheric downlink radiation amount corresponding to the ice and snow area; s23, acquiring a topographic digital elevation model corresponding to the ice and snow area, and carrying out atmospheric correction on an atmospheric visual radiation image data set of the ice and snow area based on the atmospheric uplink radiation amount and the atmospheric downlink radiation amount corresponding to the ice and snow area combined with the topographic digital elevation model so as to generate an ice and snow area radiation correction image data set; And S24, performing image geometric registration on the corresponding ice and snow area images in the ice and snow area radiation correction image data set based on target control points preset on the ground corresponding to the ice and snow area, and controlling the ice and snow pixel positioning error between the ice and snow area images to be smaller than 0.5 pixel so as to obtain an ice and snow area geometric registration image set sequence.
4. The method for identifying and extracting ice and snow interface features based on optical images as recited in claim 3, wherein the step S22 includes the steps of: Acquiring atmospheric visibility corresponding to real-time measurement; acquiring the corresponding water vapor content of real-time measurement; The method comprises the steps of carrying out atmospheric radiation influence assessment on an ice and snow area based on atmospheric visibility and water vapor content to obtain an atmospheric radiation transmission influence coefficient corresponding to the atmospheric visibility and the water vapor, carrying out three-dimensional radiation analysis modeling on an atmosphere corresponding to the ice and snow area based on the atmospheric radiation transmission influence coefficient corresponding to the atmospheric visibility and the water vapor, and incorporating the ground reflectivity, solar radiation intensity and water vapor and aerosol distribution in an atmosphere layer corresponding to the ice and snow surface into a model to generate an atmospheric radiation distribution model of the ice and snow area; Performing uplink and downlink radiation transmission simulation on the ice and snow area based on an atmospheric radiation distribution model of the ice and snow area so as to generate a corresponding radiation transmission path of the radiation energy of the atmosphere in the ice and snow area in the uplink and downlink processes; and carrying out accurate calculation on the uplink and downlink radiation of the atmosphere in the ice and snow area based on the corresponding radiation transmission paths in the uplink and downlink processes of the radiation energy of the atmosphere in the ice and snow area so as to obtain the corresponding uplink radiation and downlink radiation of the atmosphere in the ice and snow area.
5. The method for identifying and extracting the characteristics of the ice and snow interface based on the optical image according to claim 3, wherein the step S23 of performing the atmospheric correction on the air apparent radiation image dataset of the ice and snow area based on the air uplink radiation amount and the air downlink radiation amount corresponding to the ice and snow area in combination with the topographic digital elevation model comprises the following steps: Acquiring a corresponding atmospheric uplink radiation deflection angle and a corresponding atmospheric downlink radiation deflection angle through uplink and downlink radiation of the atmosphere in the ice and snow area; Based on the corresponding atmospheric uplink radiation amount and the corresponding atmospheric downlink radiation amount in the ice and snow area, and combining the corresponding atmospheric uplink radiation deflection angle and the corresponding atmospheric downlink radiation deflection angle, carrying out uplink and downlink radiation transmission length calculation to obtain the corresponding uplink radiation transmission length and downlink radiation transmission length in the ice and snow area; Determining the corresponding atmospheric radiation transmission path length according to the uplink radiation transmission length and the downlink radiation transmission length corresponding to the ice and snow area, and carrying out terrain elevation fluctuation error statistics based on the atmospheric radiation transmission path length and combining with a terrain digital elevation model to obtain the atmospheric radiation error caused by the terrain elevation fluctuation of the ice and snow area; And carrying out atmosphere correction on each ice and snow area image in the ice and snow area atmospheric apparent radiation image data set based on the atmospheric radiation error caused by the relief of the topography in the ice and snow area so as to generate an ice and snow area radiation correction image data set.
6. The method for identifying and extracting the ice and snow interface features based on the optical image as claimed in claim 1, wherein the step S3 comprises the following steps: S31, selecting a typical ice and snow sample corresponding to an ice and snow area according to a geometric registration image set sequence of the ice and snow area, and performing ice and snow microstructure analysis to obtain surface microstructure characteristics corresponding to different types of ice and snow samples, wherein the surface microstructure characteristics comprise ice crystal morphology, particle size and porosity, and the typical ice and snow samples comprise melting snow, new snow and glacier ice; step S32, measuring the corresponding ice and snow spectral reflection characteristics of different types of ice and snow samples under each wave band by synchronously using a spectrometer, and establishing an image spectral characteristic set of an ice and snow area corresponding to the different types of ice and snow samples by using the surface microstructure characteristics and the ice and snow spectral reflection characteristics; S33, carrying out spectral feature matching calculation on pixels of different areas corresponding to the images of the ice and snow areas in the ice and snow area geometric registration image set sequence based on the image spectral feature sets of the ice and snow areas corresponding to the ice and snow areas of different types of ice and snow samples so as to obtain spectral similarity between the pixels of the different areas and the spectral features of the ice and snow types in the ice and snow areas; s34, classifying ice and snow areas based on spectral similarity between pixels of different areas in the ice and snow area images and spectral features of the ice and snow types, wherein the pixels of different areas correspond to the images of the ice and snow areas in the ice and snow area geometric registration image set sequence, so as to generate an ice and snow area type classification chart, and the ice and snow area classification chart comprises ice and snow area types corresponding to a snow melting area, a new snow area and a glacier area; and step 35, performing ice and snow interface post-treatment on the ice and snow domain type classification map to obtain an ice and snow interface contour characteristic curve.
7. The method for identifying and extracting ice and snow interface features based on optical images as recited in claim 6, wherein the step S35 includes the steps of: S351, acquiring corresponding reflectivity difference values between adjacent pixels in different types of ice and snow areas under a spectrum band through an ice and snow area type classification chart; step S352, constructing a Gaussian smooth gradient field for the corresponding reflectance difference values between adjacent pixels of different types of ice and snow areas under a spectrum band so as to smoothly convert discrete reflectance difference values into corresponding reflectance difference gradients and construct an optical reflection difference gradient field corresponding to different types of ice and snow areas; Step S353, carrying out ice and snow phase change movement analysis on different types of ice and snow areas in the ice and snow area type classification map based on optical reflection difference gradient fields corresponding to the different types of ice and snow areas so as to generate dynamic phase change movement tracks corresponding to the ice and snow area interfaces along with the time; And S354, optimizing the interface profile form according to the dynamic phase change movement track corresponding to the ice and snow area interface along with the time, and optimizing the ice and snow interfaces corresponding to different types of ice and snow areas through expansion and corrosion operation so as to obtain an ice and snow interface profile characteristic curve.
8. The method for identifying and extracting the ice and snow interface features based on the optical image as claimed in claim 1, wherein the step S4 includes the steps of: S41, performing ice and snow interface image segmentation on the geometric registration image set sequence of the ice and snow area based on the ice and snow interface contour characteristic curve to generate an ice and snow image set sequence corresponding to different ice and snow interface areas, wherein the different ice and snow interface areas comprise an ice and snow interface of a snow melting area, an ice and snow interface of a new ice and snow area and an ice and snow interface of a glacier area; Step S42, acquiring the ice and snow surface temperature condition corresponding to the ice and snow interface of the snow melting area, dividing the ice and snow interface of the snow melting area into grid units, and calculating the corresponding temperature gradient vector of each grid unit in a point time interval based on the ice and snow surface temperature condition corresponding to the ice and snow interface of the snow melting area; S43, carrying out ice and snow melting migration analysis on an ice and snow image set sequence corresponding to the ice and snow interface of the snow melting area based on temperature gradient vectors corresponding to each grid unit of the ice and snow interface of the snow melting area in a certain time interval so as to obtain corresponding melting boundary migration at the ice and snow interface; S44, estimating snow depth of an ice and snow image set sequence corresponding to an ice and snow interface in a new snow area to obtain corresponding snow change depth at the ice and snow interface; S45, carrying out ice crack evolution analysis on an ice and snow image set sequence corresponding to an ice and snow interface in a glacier area to obtain the corresponding ice crack development degree at the ice and snow interface; And S46, arranging corresponding ablation boundary migration, snow cover change depth and ice crack development degree at the ice and snow interface according to a standardized format report so as to generate an ice and snow interface dynamic characteristic monitoring report.
9. The method for identifying and extracting ice and snow interface features based on optical images as recited in claim 8, wherein the step S44 includes the steps of: acquiring a corresponding new snow accumulation depth according to time change through an ice and snow image set sequence corresponding to an ice and snow interface of a new snow area; Acquiring the porosity and temperature distribution corresponding to the new snow, and performing snow depth attenuation assessment according to the porosity and temperature distribution corresponding to the new snow to obtain the snow depth attenuation rate of the new snow area; and estimating the snow change influence of the new snow accumulation depth corresponding to the ice-snow interface in the new snow area based on the snow depth attenuation rate in the new snow area so as to obtain the corresponding snow change depth at the ice-snow interface.
10. The method for identifying and extracting ice and snow interface features based on optical images as claimed in claim 8, wherein the step S45 includes the steps of: acquiring the ice crack space distribution corresponding to the ice and snow interface of the glacier region through the ice and snow image collection sequence corresponding to the ice and snow interface of the glacier region; measuring the space density and the width of the gaps according to the space distribution of the ice gaps corresponding to the ice and snow interfaces of the glacier area so as to obtain the space density and the space width of the ice gaps corresponding to the ice and snow interfaces of the glacier area; And carrying out ice crack development evolution evaluation on the ice and snow image set sequence corresponding to the ice and snow interface of the glacier region based on the ice and snow space density and the ice and snow space width corresponding to the ice and snow interface of the glacier region so as to obtain the corresponding ice and snow development degree at the ice and snow interface.

Description

Ice and snow interface characteristic identification and extraction method based on optical image Technical Field The invention relates to the technical field of image processing, in particular to an ice and snow interface characteristic identification and extraction method based on optical images. Background Along with the influence of global climate change, the monitoring and research of ice and snow environment increasingly becomes an important subject in meteorology, geography and environmental science, and the accurate identification and extraction of ice and snow interface characteristics have important significance in the fields of research of ice and snow physical process, prediction of climate change, disaster early warning and the like. At present, an ice and snow interface characteristic recognition method based on an optical image gradually becomes a research hot spot, and the optical image has the advantages of high resolution, low acquisition cost, strong spatial resolution and the like, so that the method is widely applied to monitoring of ice and snow areas, and the optical characteristics of an ice and snow interface, such as brightness, texture, color and the like, can be effectively reflected by the image acquired by the optical image, and can be further used for extracting the characteristics of ice and snow coverage, ice and snow thickness, ice and snow quality and the like. However, in the existing optical image processing method, in a complex ice and snow environment, particularly in a low-temperature, strong light reflection or cloud and fog coverage condition, a clear ice and snow interface image cannot be obtained, and dynamic changes of the ice and snow interface cannot be monitored in real time, so that the accuracy of feature identification is reduced. Disclosure of Invention Accordingly, the present invention is directed to an optical image-based method for identifying and extracting features of ice/snow interface, and an apparatus thereof. In order to achieve the above purpose, the method for identifying and extracting the ice and snow interface features based on the optical image comprises the following steps: Step S1, acquiring an optical image sequence of an ice and snow area corresponding to visible light, near infrared and short wave infrared bands through an unmanned aerial vehicle-mounted spectrum camera, synchronously recording shooting positions, attitude angles and ice and snow elevation data, and generating an optical image dataset of the ice and snow area corresponding to a space-time tag based on the shooting positions, the attitude angles and the ice and snow elevation data; S2, performing radiometric calibration and atmospheric correction on the optical image data set of the ice and snow area to generate a radiation correction image data set of the ice and snow area; Step S3, classifying the ice and snow domain types according to the geometric registration image set sequence of the ice and snow region to generate an ice and snow domain type classification map; And S4, carrying out ice and snow interface dynamic feature recognition on the geometric registration image set sequence of the ice and snow area based on the ice and snow interface contour feature curve so as to generate an ice and snow interface dynamic feature monitoring report, wherein the ice and snow interface dynamic feature monitoring report comprises corresponding ablation boundary migration, snow cover change depth and ice crack development degree at the ice and snow interface. Further, step S1 includes the steps of: s11, obtaining topographic relief data of the ice and snow area pairs, and planning and generating flight scanning tracks of the ice and snow areas corresponding to the unmanned aerial vehicle based on the topographic relief data; S12, carrying a spectrum camera on the unmanned aerial vehicle to select visible light red, green and blue wave bands sensitive to the albedo of the ice and snow, a near infrared wave band capable of reflecting the particle size of the ice and snow and a short wave infrared wave band capable of distinguishing ice and snow from water, and shooting by utilizing the spectrum camera based on the flight scanning track of the ice and snow area corresponding to the unmanned aerial vehicle to obtain an optical image sequence of the ice and snow area corresponding to the visible light, the near infrared and the short wave infrared wave bands; S13, synchronously recording shooting positions, attitude angles and ice and snow elevation data corresponding to shooting moments through integrating an inertial measurement unit and a global navigation satellite system in the unmanned aerial vehicle; And S14, taking the shooting position, the attitude angle and the ice and snow elevation data corresponding to the shooting time as the space-time tag of the optical image of the ice and snow area corresponding to the same shooting time in the optical image sequence of the ice an