CN-121256339-B - Sea area resource feature extraction method based on high-quality data set

CN121256339BCN 121256339 BCN121256339 BCN 121256339BCN-121256339-B

Abstract

The invention discloses a sea area resource feature extraction method based on a high-quality data set, which relates to the technical field of sea area resource identification and comprises the following steps of obtaining the high-quality data set of sea area resources, named as a resource data set, wherein the resource data set comprises a data set and an image set; the method comprises the steps of extracting the characteristics of a data set in a resource data set to obtain the data characteristics of sea area resources, extracting the characteristics of an image set in the resource data set to obtain the image characteristics of the sea area resources, integrating the data characteristics and the image characteristics, primarily screening the sea area resources in the ocean through the image characteristics, and verifying the sea area resources according to the data characteristics.

Inventors

ZHANG MIAO
CHEN PEIXIONG
LI ZIFENG

Assignees

自然资源部第二海洋研究所

Dates

Publication Date: 20260512
Application Date: 20251208

Claims (7)

1. The sea area resource feature extraction method based on the high-quality data set is characterized by comprising the following steps of: acquiring a high-quality data set of sea area resources, namely a resource data set, wherein the resource data set comprises a data set and an image set; extracting features of a data set in the resource data set to obtain data features of sea area resources; Extracting features of an image set in the resource data set to obtain image features of sea area resources; Integrating the data characteristics and the image characteristics, primarily screening sea area resources in the ocean through the image characteristics, and verifying the sea area resources according to the data characteristics; extracting features of the image set in the resource data set, and acquiring the image features of the sea area resource comprises the following sub-steps: Extracting features of regional images in the image set to obtain image preliminary features of sea area resources; Calibrating the preliminary features of the image based on the image of the sea surface without sea area resources to obtain image features; extracting features of regional images in the image set, and acquiring the image preliminary features of sea area resources comprises the following sub-steps: marking a pixel point of an ith row and a jth column in the area image as PT (i, j), converting the area image into a gray level image, and marking a gray level value at the PT (i, j) as GV (i, j); Acquiring the water depth in a resource data set to which an area image belongs, namely, area depth, establishing a two-dimensional coordinate system by taking the area depth as a horizontal axis and GV (i, j) as a vertical axis, namely, a resource image analysis chart, recording the GV (i, j) into the resource image analysis chart according to the area depth, and designating a coordinate point of the resource image analysis chart as a resource image analysis point; performing function regression on the resource image analysis graph, and naming a function obtained by the function regression as a depth gray scale relation function; Obtaining residual errors of the resource image analysis points and the depth gray scale relation function, and naming the residual errors as relation fluctuation values, wherein the relation fluctuation values are the preliminary characteristics of the images; calibrating the image preliminary features based on the image of the sea surface without sea area resources, and obtaining the image features comprises the following sub-steps: randomly selecting a region without sea area resources on the sea surface, namely a non-resource region, acquiring an image of the non-resource region, and namely a non-resource image; Extracting GV (i, j) of the resource-free image, marking as W (i, j), and simultaneously acquiring the water depth of the resource-free region, namely the resource-free depth; recording W (i, j) into a resource image analysis graph according to the resource-free depth, and naming coordinates formed by W (i, j) and the resource-free depth as resource-free points; Obtaining a residual error of a relation function between a non-resource point and a depth gray scale, namely a non-resource fluctuation value, obtaining a minimum value of the non-resource fluctuation value, namely a non-resource fluctuation limit value, obtaining a maximum value of the relation fluctuation value, and namely a relation fluctuation limit value; and judging whether the non-resource fluctuation limit value is larger than or equal to the relation fluctuation limit value, if so, taking the average value of the non-resource fluctuation limit value and the relation fluctuation limit value as the image characteristic, and if not, executing the image characteristic calibration scheme.
2. The sea area resource feature extraction method based on the high-quality data set according to claim 1, wherein the sea area resource is a biological resource, the resource data set is recorded with monitoring data, the monitoring data comprises a sea water temperature, a chlorophyll concentration, a salinity, a water quality, a water depth and an area image, the set of the sea water temperature, the chlorophyll concentration, the salinity, the water quality and the water depth is a data set, and the set of the area image is an image set.
3. The method for extracting the characteristics of the sea area resources based on the high-quality data set according to claim 2, wherein the step of extracting the characteristics of the data set in the resource data set and acquiring the data characteristics of the sea area resources comprises the following sub-steps: counting different monitoring data in the data set to obtain the individual range of the monitoring data; And extracting the characteristics of the individual range to obtain the data characteristics of the sea area resources.
4. A method for extracting sea resource features based on high quality data set as claimed in claim 3, wherein the statistics of different monitoring data in the data set to obtain individual range of the monitoring data comprises the following sub-steps: The method comprises the steps of (1) designating the area where biological resources are located as biological areas, wherein each biological area is provided with a resource data set, and the resource data sets of different biological areas are independent; Acquiring respective ranges of each item of monitoring data in the resource data set, naming the respective ranges as biological monitoring ranges, numbering biological areas, and marking the biological areas as BY n , wherein n is a positive integer and n is a serial number of BY; the biological monitoring range comprises a seawater temperature range, a chlorophyll concentration range, a salinity range, a water quality range and a water depth range; the seawater temperature range, chlorophyll concentration range, salinity range, water quality range and water depth range in the biological monitoring range are individual ranges; The individual ranges of BY n are labeled BF (n, m), where m is a positive integer and (n, m) is the serial number of BF, and simultaneously, BF (n, m) represents the seawater temperature range, chlorophyll concentration range, salinity range, water quality range and water depth range in BY n in the order of m from small to large, respectively.
5. The method for extracting sea resource features based on high-quality data set as claimed in claim 4, wherein the step of extracting the features of the individual range to obtain the data features of the sea resource comprises the following sub-steps: Counting the median value of each BF (n, m), marking as BG (n, m), establishing a two-dimensional coordinate system by taking m as an X axis and BG (n, m) as a Y axis, naming the coordinate system as a resource characteristic analysis chart, and recording the BG (n, m) into the resource characteristic analysis chart according to m; the coordinate points in the resource feature analysis graph are named as feature analysis points, adjacent feature analysis points are connected through straight lines to obtain feature fold lines, each n value corresponds to one feature fold line, and all the feature fold lines are in the same resource feature analysis graph; obtaining feature analysis points of the maximum value and the minimum value of the Y axis corresponding to each value of the X axis in the resource feature analysis graph, and naming the feature analysis points as the maximum feature point and the minimum feature point; And connecting the maximum characteristic point and the minimum characteristic point with the same value of the X axis through a straight line to finally form a surface, and naming the surface as a data characteristic.
6. The high quality dataset based sea area resource feature extraction method as claimed in claim 5, wherein the image feature calibration scheme comprises the sub-steps of: the method comprises the steps of (1) uniformly dividing the analysis fluctuation value into a first quantity range, naming the fluctuation range, and carrying out sequencing numbering on the fluctuation range according to a sequence from small to large, wherein the sequence is represented by a symbol H t , t is a positive integer and t is a serial number of H; Counting the number of non-resource fluctuation values in H t and the number of relation fluctuation values, wherein the number is marked as FA t and FB t respectively; Establishing histograms by taking H t as X axes and the number as Y axes, naming the histograms as characteristic calibration graphs, inputting the characteristic calibration graphs into FA t and FB t according to H t , naming a square column formed by FA t and H t as a resource-free analysis column, and naming a square column formed by FB t and H t as a relation analysis column; Continuously adding one to the first quantity to increase the quantity of the fluctuation range, enabling the final square column to approach to a smooth curve, naming the smooth curve formed by the non-resource analysis columns as a non-resource analysis curve, and naming the smooth curve formed by the relation analysis columns as a relation analysis curve; And acquiring an intersection point of the resource-free analysis curve and the relation analysis curve, namely a difference intersection point, and naming H t at the difference intersection point as an image feature.
7. The method for extracting sea area resource features based on high quality data set as claimed in claim 6, wherein integrating the data features and the image features, primarily screening sea area resources in the ocean by the image features, and verifying the sea area resources based on the data features comprises the sub-steps of: Acquiring a marine data set, wherein the marine data set records water depths of different areas in the ocean, the water depths are named as reference depths, and the areas corresponding to the reference depths are named as reference areas; Substituting the reference depth into a depth gray relation function, solving to obtain reference gray, adding the reference gray and the image characteristics to obtain a reference upper limit, and subtracting the reference gray and the image characteristics to obtain a reference lower limit; Marking pixel points with gray values between a reference lower limit and a reference upper limit in a reference area as suspected points, and naming an area formed by continuous adjacent suspected points as a suspected area; sending the unmanned ship to the suspected area to collect monitoring data, naming the monitoring data as reference data, analyzing characteristic broken lines of the reference data, and naming the characteristic broken lines as reference broken lines; substituting the reference broken line into the resource characteristic analysis chart, and marking that the biological resource exists in the reference area if the reference broken line is in the data characteristic.

Description

Sea area resource feature extraction method based on high-quality data set Technical Field The invention relates to the technical field of sea area resource identification, in particular to a sea area resource feature extraction method based on a high-quality data set. Background The sea area resource identifying technology is one of the technology of utilizing space to earth observation, sea site detection, computer science, artificial intelligence, etc. to find, locate, classify and quantify natural resource, such as fishery resource, oil and gas resource, mineral resource and spatial resource, such as port, channel, tourist area, culturable area. The existing sea area resource identification technology generally needs to detect each area in the ocean to find out biological resources in the ocean, although only offshore biological resources are needed to be detected, the offshore scope is quite large, the offshore area cannot be comprehensively detected, and therefore selective detection is needed, the existing sea area resource identification technology cannot selectively detect the ocean, the existing sea area resource identification technology mostly adopts a mode of comparing preset thresholds to identify sea area resources, the different factors in the sea area have correlations, the numerical value of one parameter is different, the applicable scope of the other parameters is also different, and the existing sea area resource identification technology also has the problem that manpower resources and material resources are excessively wasted when searching sea area resources. Disclosure of Invention The invention aims to solve at least one of the technical problems in the prior art to a certain extent, a high-quality data set of sea area resources is obtained and named as a resource data set, the resource data set comprises a data set and an image set, different monitoring data in the data set are counted to obtain individual ranges of the monitoring data, the individual ranges are subjected to feature extraction to obtain data features of the sea area resources, meanwhile, region images in the image set are subjected to feature extraction to obtain image preliminary features of the sea area resources, the image preliminary features are calibrated based on images of sea surfaces without the sea area resources to obtain image features, finally the data features and the image features are integrated, sea area resources are preliminarily screened in the sea through the image features, and verification is carried out according to the data features, so that the problem that the existing sea area resource identification technology is unreasonable in identifying the sea area resources, and human resources and material resources are excessively wasted when the sea area resources are searched is solved. In order to achieve the above object, the present application provides a sea area resource feature extraction method based on a high quality data set, comprising the steps of: acquiring a high-quality data set of sea area resources, namely a resource data set, wherein the resource data set comprises a data set and an image set; extracting features of a data set in the resource data set to obtain data features of sea area resources; Extracting features of an image set in the resource data set to obtain image features of sea area resources; Integrating the data features and the image features, primarily screening sea area resources in the ocean through the image features, and verifying the sea area resources according to the data features. Further, the sea area resource is a biological resource, the resource data set is recorded with monitoring data, the monitoring data comprise sea water temperature, chlorophyll concentration, salinity, water quality, water depth and area images, the set formed by the sea water temperature, the chlorophyll concentration, the salinity, the water quality and the water depth is a data set, and the set formed by the area images is an image set. Further, extracting features of the data set in the resource data set, and acquiring the data features of the sea area resource comprises the following sub-steps: counting different monitoring data in the data set to obtain the individual range of the monitoring data; And extracting the characteristics of the individual range to obtain the data characteristics of the sea area resources. Further, the statistics of different monitoring data in the data set is carried out, and the individual range of the monitoring data is obtained, which comprises the following substeps: The method comprises the steps of (1) designating the area where biological resources are located as biological areas, wherein each biological area is provided with a resource data set, and the resource data sets of different biological areas are independent; Acquiring respective ranges of each item of monitoring data in the resource data set, naming the respective ranges as b