CN-117194700-B - Satellite hyperspectral sample library construction method, device and equipment and readable storage medium

CN117194700BCN 117194700 BCN117194700 BCN 117194700BCN-117194700-B

Abstract

The application provides a method, a device, equipment and a readable storage medium for constructing a satellite hyperspectral sample library, which relate to the technical field of satellite hyperspectral sample libraries and comprise the steps of obtaining at least two pieces of satellite hyperspectral data and basic data corresponding to the satellite hyperspectral data, carrying out gradient segmentation according to elevation data corresponding to each piece of satellite hyperspectral data to obtain at least two pieces of elevation sub-data, carrying out weathered vegetation factor partition calculation based on each piece of satellite hyperspectral data to obtain a partition result of the satellite hyperspectral data, and marking each pixel in each piece of satellite hyperspectral data according to geological data, the elevation sub-data and the partition result to obtain marked satellite hyperspectral data, wherein each pixel in the marked satellite hyperspectral data comprises weathered vegetation, lithology types and classification attributes in three dimensions of gradient. The sample of the application meets the requirements of accurate interpretation of different levels of scenes, targets, pixels and the like.

Inventors

YUAN XIAOBO
XIE MENG
LIU HANHU
ZHANG RUI
XU ZHENGXUAN
ZHANG YINGXU
TONG PENG
MA MINGMING

Assignees

中国国家铁路集团有限公司
中铁工程设计咨询集团有限公司
中铁二院工程集团有限责任公司
成都理工大学
西南交通大学

Dates

Publication Date: 20260512
Application Date: 20230711

Claims (6)

1. The method for constructing the satellite hyperspectral sample library is characterized by comprising the following steps of: Acquiring at least two pieces of satellite hyperspectral data and basic data corresponding to the satellite hyperspectral data, wherein the basic data comprises elevation data and geological data; Slope segmentation is carried out according to elevation data corresponding to each part of satellite hyperspectral data, at least two parts of elevation sub-data are obtained, and the gradients contained in each part of elevation sub-data are different from each other; Performing weathered vegetation factor partition calculation based on each piece of satellite hyperspectral data to obtain a partition result of the satellite hyperspectral data; Labeling each pixel in each piece of satellite hyperspectral data according to the geological data, the elevation sub-data and the partitioning result to obtain labeled satellite hyperspectral data, wherein each pixel in the labeled satellite hyperspectral data comprises classification attributes in three dimensions of weathered vegetation, lithology type and gradient; Sequentially carrying out sample cutting, sample inner same attribute combination and sample coding on the marked satellite hyperspectral data, and taking the satellite hyperspectral data after sample coding as one sample in a satellite hyperspectral sample library; the method also comprises sample screening in a satellite hyperspectral sample library, wherein the sample screening comprises the following steps: All original hyperspectral curves in the sample are firstly subjected to envelope elimination, and then smoothing treatment is carried out to obtain smooth hyperspectral curves after all corresponding envelope curves in the sample are eliminated; according to the smooth curves with the envelope curves eliminated, screening to obtain absorption peak parameters corresponding to each smooth hyperspectral curve, wherein the absorption peak parameters comprise the number of absorption peaks and the wavelength of each absorption peak; according to all the preprocessed original hyperspectral curves, calculating to obtain a difference evaluation value corresponding to each smooth hyperspectral curve; Carrying out pixel point marking according to the difference evaluation value and the absorption peak parameter corresponding to each smooth hyperspectral curve; Calculating to obtain a marking rate according to marked pixel points and total pixel points in the sample; Judging according to the marking rate to obtain a screening result of the current sample; Wherein, according to all original hyperspectral curves after pretreatment, calculating to obtain a difference evaluation value corresponding to each hyperspectral curve, including: normalizing all the preprocessed original hyperspectral curves to obtain normalized hyperspectral curves corresponding to each original hyperspectral curve; constructing an evaluation matrix corresponding to each original hyperspectral curve according to all normalized hyperspectral curves, wherein each row of elements of the evaluation matrix are respectively the contrast value of one reference hyperspectral curve and one curve in all normalized hyperspectral curves, wherein the reference hyperspectral curve is one curve in all normalized hyperspectral curves, and the evaluation matrix is an asymmetric matrix; According to the average number of all elements in the evaluation matrix corresponding to each original hyperspectral curve; and calculating a difference value according to the evaluation matrix corresponding to each original hyperspectral curve, and taking the average and the difference value as difference evaluation values.
2. The method for constructing a satellite hyperspectral sample library according to claim 1, wherein the performing the division calculation of the weathered vegetation factor based on each piece of the satellite hyperspectral data to obtain the division result corresponding to the satellite hyperspectral data includes: acquiring an optical image of each satellite hyperspectral data corresponding to the same period; Calculating the vegetation index of each pixel point according to the NDVI or RENDVI calculation method; dividing the satellite hyperspectral data into four areas according to the vegetation index of each pixel point in the satellite hyperspectral image, wherein the four areas comprise vegetation-free areas; And interpreting the vegetation-free area according to the satellite hyperspectral data and the contemporaneous optical image, and dividing the vegetation-free area according to an interpretation result to obtain an original rock exposure area and a weathered object area.
3. A satellite hyperspectral sample library construction apparatus, comprising: the data acquisition unit is used for acquiring at least two pieces of satellite hyperspectral data and basic data corresponding to the satellite hyperspectral data, wherein the basic data comprises elevation data and geological data; The segmentation calculation unit is used for carrying out gradient segmentation according to the elevation data corresponding to each part of the satellite hyperspectral data to obtain at least two parts of elevation sub-data, wherein the gradients contained in each part of elevation sub-data are different; the partition calculation unit is used for carrying out partition calculation on the weathered vegetation factors based on each piece of satellite hyperspectral data to obtain partition results of the satellite hyperspectral data; The labeling unit is used for labeling each pixel in each piece of satellite hyperspectral data according to the geological data, the elevation sub-data and the partition result to obtain labeled satellite hyperspectral data, and each pixel in the labeled satellite hyperspectral data comprises classification attributes in three dimensions of weathered vegetation, lithology type and gradient; the coding unit is used for sequentially carrying out sample cutting, sample inner same attribute combination and sample coding on the marked satellite hyperspectral data, and taking the satellite hyperspectral data after sample coding as one sample in a satellite hyperspectral sample library; The satellite hyperspectral sample library constructing device further comprises a screening unit, wherein the screening unit is used for screening samples in the satellite hyperspectral sample library, and the sample screening comprises the following steps: All original hyperspectral curves in the sample are firstly subjected to envelope elimination, and then smoothing treatment is carried out to obtain smooth hyperspectral curves after all corresponding envelope curves in the sample are eliminated; according to the smooth curves with the envelope curves eliminated, screening to obtain absorption peak parameters corresponding to each smooth hyperspectral curve, wherein the absorption peak parameters comprise the number of absorption peaks and the wavelength of each absorption peak; according to all the preprocessed original hyperspectral curves, calculating to obtain a difference evaluation value corresponding to each smooth hyperspectral curve; Carrying out pixel point marking according to the difference evaluation value and the absorption peak parameter corresponding to each smooth hyperspectral curve; Calculating to obtain a marking rate according to marked pixel points and total pixel points in the sample; Judging according to the marking rate to obtain a screening result of the current sample; The satellite hyperspectral sample library construction device further comprises a correction unit, wherein the correction unit is used for correcting all satellite hyperspectral data and updating the satellite hyperspectral data into corrected data, and the correction unit comprises: the first correction unit is used for carrying out radiation correction, atmosphere correction and geometric correction on all the satellite hyperspectral data so as to obtain all the preliminarily corrected satellite hyperspectral data; The second correction unit is used for carrying out terrain radiation correction on all the satellite hyperspectral data after primary correction to obtain all the satellite hyperspectral data after secondary correction; and the third correction unit is used for carrying out error value and negative value elimination processing on all the satellite hyperspectral data after the secondary correction to obtain corrected data.
4. The apparatus for constructing a satellite hyperspectral sample library according to claim 3, wherein the partition calculating unit includes: the image acquisition unit is used for acquiring optical images corresponding to the same period of each satellite hyperspectral data; the image calculation unit is used for calculating the vegetation index of each pixel point of the satellite hyperspectral image according to the NDVI or RENDVI calculation method; the regional division unit is used for carrying out regional division on the satellite hyperspectral data according to the vegetation index of each pixel point in the satellite hyperspectral image to obtain four subareas, wherein the subareas comprise vegetation-free areas; And the depth interpretation unit is used for interpreting the vegetation-free area according to the satellite hyperspectral data and the contemporaneous optical image, and dividing the vegetation-free area according to an interpretation result to obtain an original rock exposure area and a weathered object area.
5. A satellite hyperspectral sample library construction apparatus, characterized by comprising: A memory for storing a computer program; A processor for implementing the steps of the method for constructing a satellite hyperspectral sample library as claimed in any one of claims 1 to 2 when executing the computer program.
6. A readable storage medium, characterized in that it has stored thereon a computer program which, when executed by a processor, implements the steps of the method for constructing a satellite hyperspectral sample library according to any one of claims 1 to 2.

Description

Satellite hyperspectral sample library construction method, device and equipment and readable storage medium Technical Field The invention relates to the technical field of hyperspectral image processing, in particular to a method, a device and equipment for constructing a satellite hyperspectral sample library and a readable storage medium. Background The existing remote sensing interpretation sample library mainly comprises a spectrum curve sample set, RGB (visible light) images and the like, but the sample set has different maps, a non-uniform classification system and lacks a hyperspectral lithology sample library aiming at engineering geological rock groups. The number and types of the existing deep learning samples are limited, and the multi-scale, multi-sensor and multi-phase performance of the remote sensing image are not enough, so that an engineering geological rock satellite hyperspectral sample library oriented to intelligent interpretation is needed at present so as to improve the intelligent interpretation accuracy of engineering geological lithology interpretation. Disclosure of Invention The invention aims to provide a satellite hyperspectral sample library construction method, device and equipment and a readable storage medium, so as to solve the problems. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: The application provides a satellite hyperspectral sample library construction method, which comprises the steps of obtaining at least two pieces of satellite hyperspectral data and basic data corresponding to the satellite hyperspectral data, wherein the basic data comprise elevation data and geological data, carrying out gradient segmentation according to the elevation data corresponding to each piece of satellite hyperspectral data to obtain at least two pieces of elevation sub-data, carrying out weathered vegetation factor partition calculation according to each piece of satellite hyperspectral data to obtain a partition result of the satellite hyperspectral data, marking each pixel in each piece of satellite hyperspectral data according to the geological data, the elevation sub-data and the partition result, obtaining marked satellite hyperspectral data, wherein each pixel in the marked satellite hyperspectral data comprises weathered vegetation, lithology type and classification attribute in three dimensions, sequentially carrying out sample cutting, sample inner identical attribute merging and sample encoding on the marked satellite hyperspectral data, and taking the sample encoded satellite hyperspectral data as a sample in the hyperspectral sample library of the satellite hyperspectral data. The application further provides a satellite hyperspectral sample library construction device, which comprises a data acquisition unit, a segmentation calculation unit and a coding unit, wherein the data acquisition unit is used for acquiring at least two pieces of satellite hyperspectral data and basic data corresponding to the satellite hyperspectral data, the basic data comprise elevation data and geological data, the segmentation calculation unit is used for carrying out gradient segmentation according to the elevation data corresponding to each piece of satellite hyperspectral data to obtain at least two pieces of elevation sub-data, gradients contained in each piece of elevation sub-data are different from each other, the segmentation calculation unit is used for carrying out weathered vegetation factor partition calculation based on each piece of satellite hyperspectral data to obtain a partition result of the satellite hyperspectral data, the labeling unit is used for labeling each pixel in each piece of satellite hyperspectral data according to the geological data, the elevation sub-data and the partition result, each pixel point in the labeled satellite hyperspectral data comprises a classification vegetation, a lithology type and three dimensions, the coding unit is used for carrying out the same-grade and coding of a sample in sequence and taking the satellite hyperspectral sample as a sample in a hyperspectral sample library after the satellite hyperspectral sample is cut. In a third aspect, the present application also provides a satellite hyperspectral sample library construction apparatus, including: A memory for storing a computer program; And the processor is used for realizing the steps of the satellite hyperspectral sample library construction method when executing the computer program. In a fourth aspect, the present application also provides a readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the above-described satellite-based hyperspectral sample library construction method. The beneficial effects of the invention are as follows: the sample provided by the application is a comprehensive sample integrating scenes, targets and pixels, has the advanta