CN-121999372-A - Mariculture identification method based on active and passive remote sensing band coupling color characteristics

CN121999372ACN 121999372 ACN121999372 ACN 121999372ACN-121999372-A

Abstract

The application relates to a marine culture identification method based on active and passive remote sensing band coupling color characteristics. The method comprises the steps of obtaining optical remote sensing data and radar remote sensing data of a sea water area to be identified, carrying out feature extraction on the radar remote sensing data to obtain structural feature information, carrying out feature extraction on the optical remote sensing data to obtain spectral feature information, mapping the structural feature information and the spectral feature information into a CIE color space to obtain CIE color index features, carrying out classification prediction based on the CIE color index features, and dividing the sea water area to be identified into a sea water aquaculture area and a non-aquaculture area. By adopting the method, the detection accuracy of the mariculture area can be improved.

Inventors

HOU XUEJIAO
WANG CHENGLONG
CHEN JINGMAN
CHENG XIAO
WANG TIANXING

Assignees

中山大学

Dates

Publication Date: 20260508
Application Date: 20260408

Claims (10)

1. The marine culture identification method based on the active and passive remote sensing band coupling color characteristics is characterized by comprising the following steps of: acquiring optical remote sensing data and radar remote sensing data of a sea water area to be identified; extracting features of the radar remote sensing data to obtain structural feature information; extracting the characteristics of the optical remote sensing data to obtain spectral characteristic information; Mapping the structural feature information and the spectral feature information to a CIE color space to obtain CIE color index features; and classifying and predicting based on the CIE color index characteristics, and dividing the seawater area to be identified into a seawater culture area and a non-culture area.
2. The method according to claim 1, wherein the performing feature extraction on the radar remote sensing data to obtain structural feature information includes: performing first preprocessing on the radar remote sensing data to obtain a back scattering energy image; And extracting the backscattering coefficient of each polarized wave band from the backscattering energy image as structural characteristic information.
3. The method of claim 1, wherein the performing feature extraction on the optical remote sensing data to obtain spectral feature information comprises: Performing second preprocessing on the optical remote sensing data to obtain an earth surface reflectivity image; and extracting the reflectivity of the surface reflectivity image to obtain the reflectivity value of each spectrum band, and taking the reflectivity value as spectrum characteristic information.
4. The method of claim 1, wherein mapping the structural feature information and the spectral feature information to a CIE color space results in CIE color index characteristics, comprising: Determining a first backscattering coefficient of a vertical transmission vertical reception polarization band and a second backscattering coefficient of the vertical transmission horizontal reception polarization band according to the structural feature information; Determining the reflectivity ratio of the red-green wave band according to the structural characteristic information; and carrying out nonlinear fusion on the first backscattering coefficient, the second backscattering coefficient and the reflectivity ratio to obtain CIE color index characteristics.
5. The method of claim 4, wherein determining the reflectance ratio of the red and green bands based on the structural feature information comprises: According to the structural feature information, determining a first reflectivity of a red wave band and a second reflectivity of a green wave band; and determining the ratio of the first reflectivity to the second reflectivity as the ratio of the reflectivities of the red and green wave bands.
6. The method according to any one of claims 1 to 5, wherein the classifying prediction based on the CIE color index characteristics, dividing the seawater area to be identified into a mariculture area and a non-culture area, comprises: selecting a reflectivity value of a target spectrum band from the spectrum characteristic information, wherein the target spectrum band is determined according to the characteristic importance of each spectrum band in the spectrum characteristic information; inputting the CIE color index characteristics and the reflectivity value of the target spectrum band into a target classification model to obtain classification results of pixels in the sea water area to be identified; Dividing the seawater area to be identified according to the classification result of each pixel to obtain a seawater culture area and a non-culture area.
7. The method according to claim 6, wherein the dividing the sea water area to be identified according to the classification result of each pixel to obtain a sea water cultivation area and a non-cultivation area comprises: based on the classification result of each pixel, sequentially performing open operation processing and close operation processing on the seawater area to be identified by using basic structure elements to obtain a morphological noise reduction image; And carrying out connected domain filtering on the morphological noise reduction image to obtain a mariculture area and a non-culture area.
8. The method of claim 6, wherein the method further comprises: acquiring an optical data sample and a radar data sample; extracting features of the radar data sample to obtain a structural feature sample; Extracting the characteristics of the optical data sample to obtain a spectrum characteristic sample; mapping the structural feature sample and the spectral feature sample to a CIE color space to obtain a CIE color feature sample; according to the spectrum characteristic sample, determining the reflectivity value of each spectrum band; training an initial tree model based on the CIE color feature samples and reflectivity values of the spectral bands to obtain a target tree model; performing feature importance analysis on the target tree model to obtain feature importance values of the spectral bands; and screening out a target spectrum band from each spectrum band according to the characteristic importance value.
9. A marine culture identification device, the device comprising: The data acquisition module is used for acquiring optical remote sensing data and radar remote sensing data of the sea water area to be identified; the radar feature extraction module is used for carrying out feature extraction on the radar remote sensing data to obtain structural feature information; The optical characteristic extraction module is used for carrying out characteristic extraction on the optical remote sensing data to obtain spectral characteristic information; The feature mapping module is used for mapping the structural feature information and the spectral feature information to a CIE color space to obtain CIE color index features; And the classification prediction module is used for performing classification prediction based on the CIE color index characteristics and dividing the seawater area to be identified into a seawater culture area and a non-culture area.
10. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 8 when the computer program is executed.

Description

Mariculture identification method based on active and passive remote sensing band coupling color characteristics Technical Field The application relates to the technical field of mariculture identification, in particular to a mariculture identification method based on active and passive remote sensing wave band coupling color characteristics. Background Along with the development of coastal economy, the scale of mariculture (such as net cage and floating raft culture) is continuously enlarged. Although mariculture provides rich seafood, unordered expansion also brings problems of marine ecological environment damage, channel blockage and the like. Therefore, rapid and accurate monitoring of the distribution of mariculture areas is critical for ocean space planning and ecological protection. At present, the distribution identification of the mariculture area is realized mainly by utilizing the optical characteristics of culture facilities (algae are usually attached), however, the culture facilities in partial sea areas are not obvious in characteristics of true color images, NDVI (Normalized Difference Vegetation Index, normalized vegetation index), FAI (Floating Algae Index ) and other common optical indexes, and the difference between the culture facilities and background seawater is tiny, so that missed detection is extremely easy to cause, and the detection accuracy of the mariculture area is lower. Disclosure of Invention Based on the above, it is necessary to provide a mariculture identification method based on active and passive remote sensing band coupled color features, which can improve the detection accuracy of a mariculture area. In a first aspect, the application provides a marine culture identification method based on active and passive remote sensing band coupling color characteristics, comprising the following steps: acquiring optical remote sensing data and radar remote sensing data of a sea water area to be identified; extracting features of the radar remote sensing data to obtain structural feature information; extracting the characteristics of the optical remote sensing data to obtain spectral characteristic information; mapping the structural feature information and the spectral feature information to a CIE color space to obtain CIE color index features; And classifying and predicting based on CIE color index characteristics, and dividing the seawater area to be identified into a seawater culture area and a non-culture area. In one embodiment, feature extraction is performed on radar remote sensing data to obtain structural feature information, including: Performing first preprocessing on the radar remote sensing data to obtain a back scattering energy image; and extracting the backscattering coefficient of each polarized wave band from the backscattering energy image as structural characteristic information. In one embodiment, the feature extraction of the optical remote sensing data to obtain spectral feature information includes: performing second preprocessing on the optical remote sensing data to obtain an earth surface reflectivity image; And extracting the reflectivity of the surface reflectivity image to obtain the reflectivity value of each spectrum band, and taking the reflectivity value as spectrum characteristic information. In one embodiment, mapping the structural feature information and the spectral feature information to a CIE color space to obtain CIE color index features includes: determining a first backscattering coefficient of a vertical transmission vertical reception polarization band and a second backscattering coefficient of the vertical transmission horizontal reception polarization band according to the structural feature information; determining the reflectivity ratio of the red-green wave band according to the structural feature information; And carrying out nonlinear fusion on the first backscattering coefficient, the second backscattering coefficient and the reflectance ratio to obtain CIE color index characteristics. In one embodiment, determining the reflectance ratio of the red and green bands according to the structural feature information includes: Determining a first reflectivity of a red wave band and a second reflectivity of a green wave band according to the structural feature information; the ratio of the first reflectance to the second reflectance is determined as the reflectance ratio of the red-green band. In one embodiment, classification prediction is performed based on CIE color index characteristics, and the sea water area to be identified is divided into a sea water cultivation area and a non-cultivation area, including: Selecting a reflectivity value of a target spectrum band from the spectrum characteristic information, wherein the target spectrum band is determined according to the characteristic importance of each spectrum band in the spectrum characteristic information; Inputting CIE color index features and reflectivity values of ta