CN115639354B - Marine plastic identification method and device - Google Patents
Marine plastic identification method and device Download PDFInfo
- Publication number
- CN115639354B CN115639354B CN202211645479.5A CN202211645479A CN115639354B CN 115639354 B CN115639354 B CN 115639354B CN 202211645479 A CN202211645479 A CN 202211645479A CN 115639354 B CN115639354 B CN 115639354B
- Authority
- CN
- China
- Prior art keywords
- marine
- plastic
- identified
- sample
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a method and a device for identifying marine plastics, wherein the method comprises the following steps: acquiring corresponding ocean plastic similarity, ocean plastic map matching degree and element content according to a plastic sample to be identified; according to the similarity of the marine plastics, the matching degree of the marine plastic spectrum and the element content, when the plastic sample to be identified is primarily judged to be non-marine plastics by combining a KNN algorithm, acquiring the total content of marine bacteria of the plastic sample to be identified; calculating an attribution data point of the total content of the marine bacteria according to the total content of the marine bacteria and a KNN algorithm; and when the number of the data points belonging to the marine plastics is greater than that of the data points belonging to the non-marine plastics in the belonging data points of the total content of the marine bacteria, determining that the plastic sample to be identified is the marine plastics. The method and the device for identifying the marine plastics improve the identification precision of the marine plastics.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a method and a device for identifying marine plastics.
Background
The plastic garbage in the ocean is only a resource with a misplaced position, and the organic cleaning of the ocean plastics can not only prevent the plastics from forming micro-plastic pollution in the ocean, but also protect the ecology; in addition, the organic recycled plastic garbage is recycled, so that the sources of renewable resources are expanded, and carbon emission caused by direct incineration after the plastic garbage is salvaged ashore can be reduced.
Because of the great difference between marine environment and land environment, there are some differences between marine plastics obtained from marine environment and plastics obtained from land in terms of physical and chemical properties and surface attachments. For example, marine plastics are subject to the constant action of the marine environment, marine fouling organisms, marine bacteria, etc., leaving characteristic features on the surface and causing compositional changes.
The existing marine plastic identification method only adopts a manual visual inspection mode to identify, so that the identification precision of the marine plastic is not high.
Disclosure of Invention
The embodiment of the invention provides a method and a device for identifying marine plastics, which improve the identification precision of the marine plastics.
A first aspect of an embodiment of the present application provides a method for identifying marine plastics, including:
acquiring corresponding ocean plastic similarity, ocean plastic map matching degree and element content according to a plastic sample to be identified;
according to the similarity of the marine plastics, the matching degree of the marine plastic spectrum and the element content, when the plastic sample to be identified is primarily judged to be non-marine plastics by combining a KNN algorithm, acquiring the total content of marine bacteria of the plastic sample to be identified;
calculating an attribution data point of the total content of the marine bacteria according to the total content of the marine bacteria and a KNN algorithm;
and when the number of the data points belonging to the marine plastics is greater than that of the data points belonging to the non-marine plastics in the belonging data points of the total content of the marine bacteria, determining that the plastic sample to be identified is the marine plastics.
In a possible implementation manner of the first aspect, the method for primarily determining that the plastic sample to be identified is non-marine plastic by combining the KNN algorithm according to the marine plastic similarity, the marine plastic spectrum matching degree and the element content specifically comprises the following steps:
after normalization processing is carried out according to test data in the marine plastic database and the non-marine plastic database, a first characteristic space is established;
performing normalization processing according to the similarity of the marine plastics, the matching degree of the marine plastic map and the element content to serve as a first data sample to be identified;
calculating a Euclidean distance between the first data sample to be identified and the first feature space according to the KNN algorithm, and calculating an attribution data point of the first data sample to be identified according to the Euclidean distance;
and when the number of the data points attributed to the non-marine plastics in the attribution data points of the first data sample to be identified is larger than that of the data points attributed to the marine plastics, the plastic sample to be identified is judged to be the non-marine plastics for the first time.
In a possible implementation manner of the first aspect, the attribution data point of the total content of marine bacteria is calculated according to the total content of marine bacteria and the KNN algorithm, specifically:
performing normalization processing according to the total content data of the marine bacteria in the marine plastic database and the non-marine plastic database, and establishing a second characteristic space;
taking the total content of marine bacteria as a second data sample to be identified;
and calculating the Euclidean distance between the second data sample to be identified and the second feature space according to the KNN algorithm, and calculating the attribution data point of the total content of the marine bacteria according to the Euclidean distance.
In a possible implementation manner of the first aspect, the obtaining of the corresponding marine plastic similarity according to the plastic sample to be identified specifically includes:
and acquiring the characteristic morphology of the plastic sample to be identified, comparing the characteristic morphology of the plastic sample to be identified with data in a preset marine plastic characteristic morphology database, calculating the corresponding marine plastic similarity of the plastic sample to be identified according to the characteristic morphology comparison result, and acquiring the corresponding marine plastic similarity.
In a possible implementation manner of the first aspect, the method further includes:
and acquiring the surface adsorbate infrared spectrum of the plastic sample to be identified, comparing the surface adsorbate infrared spectrum of the plastic sample to be identified with data in a preset marine plastic infrared spectrum database, and calculating according to the infrared spectrum comparison result to obtain the marine plastic spectrum matching degree corresponding to the plastic sample to be identified.
In a possible implementation manner of the first aspect, the method further includes:
and obtaining the surface adsorbate of the plastic sample to be identified, carrying out SEM-EDS test according to the surface adsorbate, and obtaining the surface element content of the plastic sample to be identified according to the test result.
A second aspect of embodiments of the present application provides a marine plastic identification device, comprising: the device comprises an acquisition module, a first judgment module, a calculation module and a second judgment module;
the acquisition module is used for acquiring corresponding marine plastic similarity, marine plastic map matching degree and element content according to the plastic sample to be identified;
the first judgment module is used for acquiring the total content of marine bacteria of the plastic sample to be identified when the plastic sample to be identified is primarily judged to be non-marine plastic by combining a KNN algorithm according to the similarity of the marine plastics, the matching degree of a marine plastic map and the element content;
the calculation module is used for calculating attribution data points of the total content of the marine bacteria according to the total content of the marine bacteria and a KNN algorithm;
the second judging module is used for judging that the plastic sample to be identified is the marine plastic when the number of the data points belonging to the marine plastic is larger than that of the data points belonging to the non-marine plastic in the belonging data points of the total content of the marine bacteria.
In a possible implementation manner of the second aspect, the method for primarily judging that the plastic sample to be identified is non-marine plastic by combining the KNN algorithm according to the marine plastic similarity, the marine plastic map matching degree and the element content specifically comprises the following steps:
after normalization processing is carried out according to test data in the marine plastic database and the non-marine plastic database, a first characteristic space is established;
performing normalization processing according to the similarity of the marine plastics, the matching degree of the marine plastic map and the element content to serve as a first data sample to be identified;
according to the KNN algorithm, calculating the Euclidean distance between the first data sample to be identified and the first feature space, and calculating the attribution data point of the first data sample to be identified according to the Euclidean distance;
and when the number of the data points attributed to the non-marine plastics in the attribution data points of the first data sample to be identified is larger than that of the data points attributed to the marine plastics, the plastic sample to be identified is judged to be the non-marine plastics for the first time.
In a possible implementation manner of the second aspect, the attribution data points of the total content of the marine bacteria are calculated according to the total content of the marine bacteria and a KNN algorithm, and specifically are as follows:
after normalization processing is carried out according to the total marine bacteria content data in the marine plastic database and the non-marine plastic database, a second characteristic space is established;
taking the total content of marine bacteria as a second data sample to be identified;
and calculating the Euclidean distance between the second data sample to be identified and the second feature space according to the KNN algorithm, and calculating the attribution data point of the total content of the marine bacteria according to the Euclidean distance.
In a possible implementation manner of the second aspect, the obtaining of the corresponding marine plastic similarity according to the plastic sample to be identified specifically includes:
and acquiring the characteristic morphology of the plastic sample to be identified, comparing the characteristic morphology of the plastic sample to be identified with data in a preset marine plastic characteristic morphology database, calculating the corresponding marine plastic similarity of the plastic sample to be identified according to the characteristic morphology comparison result, and acquiring the corresponding marine plastic similarity.
Compared with the prior art, the method and the device for identifying the marine plastic provided by the invention comprise the following steps: acquiring corresponding ocean plastic similarity, ocean plastic map matching degree and element content according to a plastic sample to be identified; according to the similarity of the marine plastics, the matching degree of the marine plastic spectrum and the element content, when the plastic sample to be identified is primarily judged to be non-marine plastics by combining a KNN algorithm, acquiring the total content of marine bacteria of the plastic sample to be identified; calculating an attribution data point of the total content of the marine bacteria according to the total content of the marine bacteria and a KNN algorithm; and when the number of the data points belonging to the marine plastics is greater than that of the data points belonging to the non-marine plastics in the belonging data points of the total content of the marine bacteria, determining that the plastic sample to be identified is the marine plastics.
The beneficial effects are that: according to the method, when the plastic sample to be identified is primarily judged to be non-marine plastic by combining a KNN algorithm according to the similarity of the marine plastic, the matching degree of the marine plastic spectrum and the element content, whether the plastic sample to be identified is marine plastic is further identified according to the total content of marine bacteria and the KNN algorithm. According to the method, when the plastic sample to be identified is judged to be non-marine plastic for the first time, the plastic sample to be identified is further judged according to the algorithm and the total content of marine bacteria, so that the problem of misjudgment caused by judgment only according to the similarity of the marine plastic, the matching degree of a marine plastic map and the content of elements is avoided, and the accuracy of identifying the marine plastic is improved; meanwhile, the data of the plastic sample to be identified are automatically identified and classified through the algorithm, so that the problem of low identification precision caused by a manual visual inspection method in the prior art can be solved, and the identification precision and the identification efficiency of the marine plastic are greatly improved.
Furthermore, in the process of obtaining the ocean plastic similarity, the ocean plastic map matching degree, the element content and the total ocean bacteria content of the plastic sample to be identified, the method and the device can detect the plastic sample to be identified step by step according to different characteristics of the plastic sample to be identified, so that the detection efficiency and the result accuracy are improved, and meanwhile, the detection cost can be reduced.
Drawings
FIG. 1 is a schematic flow chart of a marine plastic identification method according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a first plastic sample provided in accordance with an embodiment of the present invention;
FIG. 3 is a schematic diagram of a second plastic sample provided in accordance with an embodiment of the present invention;
FIG. 4 is a schematic view of a plastic sample III provided in accordance with an embodiment of the present invention;
FIG. 5 is a schematic view of a plastic sample four provided in accordance with an embodiment of the present invention;
FIG. 6 is a schematic diagram of a feature profile of a plastic sample four provided by an embodiment of the invention;
FIG. 7 is a schematic diagram of a surface adsorbate sampling location of a plastic sample four according to an embodiment of the invention;
fig. 8 is a schematic structural diagram of a marine plastic identification device according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1, which is a schematic flow chart of a marine plastic identification method according to an embodiment of the present invention, including S101-S104:
s101: and acquiring corresponding ocean plastic similarity, ocean plastic map matching degree and element content according to the plastic sample to be identified.
In this embodiment, the obtaining of the corresponding ocean plastic similarity according to the plastic sample to be identified specifically includes:
and acquiring the characteristic morphology of the plastic sample to be identified, comparing the characteristic morphology of the plastic sample to be identified with data in a preset marine plastic characteristic morphology database, and calculating and acquiring the corresponding marine plastic similarity of the plastic sample to be identified according to a characteristic morphology comparison result.
In a specific embodiment, the method further comprises:
and acquiring the surface adsorbate infrared spectrum of the plastic sample to be identified, comparing the surface adsorbate infrared spectrum of the plastic sample to be identified with data in a preset marine plastic infrared spectrum database, and calculating according to the infrared spectrum comparison result to obtain the marine plastic spectrum matching degree corresponding to the plastic sample to be identified.
In a specific embodiment, the method further comprises:
and obtaining the surface adsorbate of the plastic sample to be identified, carrying out SEM-EDS test according to the surface adsorbate, and obtaining the surface element content of the plastic sample to be identified according to the test result.
S102: and acquiring the total content of marine bacteria of the plastic sample to be identified when the plastic sample to be identified is primarily judged to be non-marine plastic by combining a KNN algorithm according to the similarity of the marine plastic, the matching degree of the marine plastic spectrum and the element content.
In this embodiment, the primary determination that the plastic sample to be identified is non-marine plastic by combining the KNN algorithm according to the marine plastic similarity, the marine plastic map matching degree, and the element content specifically is as follows:
after normalization processing is carried out according to the test data in the marine plastic database and the non-marine plastic database, a first characteristic space is established;
performing normalization processing according to the ocean plastic similarity, the ocean plastic map matching degree and the element content to serve as a first data sample to be identified;
calculating a Euclidean distance between the first to-be-identified data sample and the first feature space according to a KNN algorithm, and calculating an attribution data point of the first to-be-identified data sample according to the Euclidean distance;
and when the number of the data points attributed to the non-marine plastics in the attributed data points of the first to-be-identified data sample is greater than that of the data points attributed to the marine plastics, determining that the to-be-identified plastic sample is the non-marine plastics for the first time.
S103: and calculating attribution data points of the total content of the marine bacteria according to the total content of the marine bacteria and a KNN algorithm.
In this embodiment, the calculating the attribution data point of the total content of the marine bacteria according to the total content of the marine bacteria and the KNN algorithm specifically includes:
performing normalization processing according to the total content data of the marine bacteria in the marine plastic database and the non-marine plastic database, and establishing a second characteristic space;
taking the total content of the marine bacteria as a second data sample to be identified;
and calculating the Euclidean distance between the second data sample to be identified and the second feature space according to a KNN algorithm, and calculating the attribution data point of the total content of the marine bacteria according to the Euclidean distance.
S104: and when the number of the data points belonging to the marine plastics is greater than that of the data points belonging to the non-marine plastics in the belonging data points of the total content of the marine bacteria, determining that the plastic sample to be identified is the marine plastics.
Further, the second feature space comprises marine bacteria total content data in a marine plastic database and marine bacteria total content data in a non-marine plastic database, K attribution data points which are nearest to the plastic sample to be identified (Euclidean distance is nearest) are obtained from the second feature space by a KNN algorithm, and as long as the Euclidean distance meets the condition, the marine bacteria total content data in the marine plastic database can be used as the attribution data points, and the marine bacteria total content data in the non-marine plastic database can also be used as the attribution data points. When the total content data of the marine bacteria in the marine plastic database is taken as an attribution data point, the attribution data point is equal to a data point which is attributed to the marine plastic; when the total marine bacteria content data in the non-marine plastic database is taken as an attribution data point, the attribution data point is equivalent to the data point attributed to the non-marine plastic.
In one embodiment, the following preparation is required during the marine plastic identification method:
1. establishing a database:
collecting the marine plastics on site, acquiring the performance parameter information of the marine plastics and the surface substances thereof, and establishing a preset marine plastic database. The performance parameter information includes surface morphology of the marine plastic and its attachments, infrared spectrum data, element composition, microorganism species distribution, etc., and the adopted equipment can be an optical microscope, a Scanning Electron Microscope (SEM), a microscopic infrared spectrometer, an X-ray fluorescence spectrometer (XRF), an Ion Chromatograph (IC), a high-throughput sequencer, etc., but is not limited thereto, and can be applied to equipment capable of analyzing the marine plastic surface morphology, infrared spectrum data, element composition, microorganism species, etc.
Accordingly, the preset marine plastic database comprises: a preset marine plastic characteristic appearance database, a preset marine plastic biological genetic information database, a preset marine plastic characteristic element database, a preset marine plastic infrared spectrum database and the like. The process of establishing the preset marine plastic database comprises the following steps: selecting characteristic features to train based on a neural network algorithm according to the shot marine plastic micro-feature picture, and establishing a preset marine plastic characteristic feature database; acquiring biological information of the surface of the marine plastic by adopting a high-throughput sequencing means, and establishing a preset marine plastic biological genetic information database; testing the content of halogen and metal elements of the marine plastic, and establishing a preset marine plastic characteristic element database; acquiring the infrared spectrum of the substances attached to the surface of the marine plastic, and establishing a preset marine plastic infrared spectrum database and the like.
Non-marine plastic data were obtained in the same manner as a control group and a non-marine plastic database was established. These non-marine plastics come from various plastics that are recycled after consumption, such as plastic bottles, plastic films, packaging bags, ropes, plastic foams, and the like.
2. Sampling method and pretreatment:
(1) The sampling process is required before the sample is granulated, for example, the sampled and inspected sample can be sourced from any stage before granulation, such as fishing, transportation, storage, sorting, cleaning, crushing and the like, but not from the stage after melting and granulation.
(2) The sampling method comprises the following steps: if the samples are in the same batch, extracting products with preset sample volume by adopting a simple random sampling mode; if the sample sources are different batches, a layered sampling mode is adopted, and the sample amount sampled in each batch is consistent with the sample amount proportion of the batch.
(3) The sampling method comprises the following steps: and (3) regarding each complete sample in the step (2) as a basic test unit. For different test items, the following specifications were followed when sampling the test unit:
1) For samples used for testing optical microscopes, scanning electron microscopes, infrared spectrums and the like, before sampling, appearance inspection and judgment are carried out on the samples, and sampling is carried out at a specified position.
2) For samples used for elemental analysis such as ion chromatography, X-ray fluorescence, ICP-MS, the whole sample can be cut into uniformly sized pieces and a specified number of samples can be obtained using the quartering method. If necessary, enough representative samples can be obtained at the designated positions of the complete samples, and then a specific number of samples meeting the test requirements can be obtained by adopting a quartering method.
3) The sampling tool needs to be sterilized in advance when the biological sample is sampled. If the size of the sample meets the requirement, the whole sample can be preferentially selected at the sampling position; if it is difficult to cover all the parts of the sample, the sample is taken at least at four different representative parts of the sample (e.g., top and bottom, front and back).
(4) Pretreatment: the sample for sampling inspection can be washed by deionized water to remove surface dirt, and can also be not cleaned.
3. Identifying the object to be tested:
step (1) visual method:
and if the surface of the sample has obvious shells attached or left by the marine organisms such as barnacles, tubificas, oysters, corals, bryozoans and the like, the sample is identified as the marine plastic. And (3) if the judgment cannot be carried out by a visual method, entering the step (2) for testing.
Step (2) an instrument analysis method:
1) The microscopic morphology analysis method comprises the following steps:
and (3) placing the plastic sample to be identified under an optical microscope to shoot the morphology, or taking a sample and then shooting the morphology by using a scanning electron microscope, selecting the characteristic morphology (namely obtaining the characteristic morphology of the plastic sample to be identified), comparing the characteristic morphology with a preset marine plastic characteristic morphology database, and outputting a similarity comparison result (namely the marine plastic similarity).
2) The microscopic infrared test method comprises the following steps:
sampling the surface adsorbate of the plastic sample to be identified, carrying out microscopic infrared test (namely obtaining the surface adsorbate infrared spectrum of the plastic sample to be identified), utilizing a preset marine plastic infrared spectrum database to carry out retrieval, and outputting a spectrum matching degree result (namely the marine plastic spectrum matching degree).
3) SEM-EDS test method:
and (3) sampling the adsorbate on the surface of the plastic sample to be identified (namely obtaining the adsorbate on the surface of the plastic sample to be identified), then carrying out SEM-EDS test, and recording the contents of elements such as C, O, ca, na, cl and the like (namely the contents of the surface elements) in the surface element composition. Further, the recorded surface element content preferably comprises: ca. Elemental contents of Na and Cl.
4) The element content testing method comprises the following steps:
selecting a proper test method according to actual needs to test the content of the elements in the sample, such as:
and sampling and shearing a plastic sample to be identified, testing the element composition of the sample by using XRF, and performing element quantification by using a self-built marking line to test the contents of F, cl and Br elements.
Sampling and digesting a plastic sample to be identified, testing halogen in the sample by using an ion chromatography, and carrying out element quantification by using a self-built marking line to test the contents of F, cl and Br elements.
In a preferred embodiment, after a plastic sample to be identified is sampled and cut into pieces, XRF or ion chromatography is used for testing halogen of the sample, a self-built marking line is used for element quantification, and only Cl element is preferably tested.
And testing the metal elements contained in the plastic sample to be identified by XRF or SEM-EDS, or testing the metal elements contained in the plastic sample by ICP-OES or ICP-MS after digestion, and performing element quantification by using the self-built marking line.
Wherein, the step 3) and the step 4) are specific processes for calculating the element content of the plastic sample to be identified.
5) Modeling and result judgment are carried out based on a KNN algorithm:
and taking various test data in the marine plastic database and the non-marine plastic database as data samples, and carrying out normalization processing on the data to establish a first characteristic space. Wherein, the mode of the most value normalization can be adopted, and X is used 1 =（X-Y min ）/（Y max -Y min ) Number of tests on each itemAnd carrying out normalization processing. Wherein X 1 For normalized data, Y max For the maximum value, Y, of this type of test data in the established feature space min Is the minimum value of the test data in the established feature space.
And (3) taking the various test results (namely the marine plastic similarity, the marine plastic map matching degree and the element content) of the plastic sample to be identified in the step (2) as characteristic parameters to carry out data extraction and normalization processing, and establishing a first data sample to be identified.
And calculating the Euclidean distance between the first data sample to be identified and each data in the feature space, selecting K data points closest to the Euclidean distance of the first data sample to be identified, counting the proportion of the data belonging to the marine plastics and the non-marine plastics, and taking the higher proportion as the class to which the first data sample to be identified belongs. The method comprises the following specific steps: when the number of the data points of the angelica which belong to the marine plastic is larger than that of the data points which belong to the non-marine plastic, the category to which the first sample to be identified belongs is the marine plastic; and when the data points of the angelica which belong to the non-marine plastics are larger than the data points which belong to the marine plastics, the class to which the first sample to be identified belongs is the non-marine plastics.
Further, in order to avoid the situation that the number of adjacent data points is equal, the K value is odd, and the K value is optimized by adopting a cross validation method.
And (4) if the category to which the first sample to be identified belongs is non-marine plastics through algorithm calculation, entering the step (3).
Step (3) a high-throughput sequencing method:
extracting a biological sample on the surface of a plastic sample to be identified by using a sampling swab, performing high-throughput sequencing and species information annotation after DNA extraction, purification and amplification, analyzing and counting biological genetic information in the sample, performing bacterial species group marking on the genus level by comparing with a Silva database, recording the information of each genus and the proportion of the genus in the population after removing unclassified bacteria, and sequentially arranging from high to low after normalization. And (3) performing matching retrieval on 95% of the self-built marine plastic biological genetic information database belonging to the category of the plastic sample to be identified, and counting the total content of marine bacteria (namely obtaining the total content of the marine bacteria of the plastic sample to be identified).
Establishing a second characteristic space by taking the total content of marine bacteria of each sample in the marine plastic database and the total content of marine bacteria of each sample in the non-marine plastic database as data samples, taking the total content of marine bacteria of the plastic sample to be identified as a data sample to be identified, obtaining K data points nearest to (Euclidean distance from) the plastic sample to be identified by a KNN algorithm, and taking the category of the data point with higher proportion as the category to which the plastic sample to be identified belongs. The method specifically comprises the following steps: when the number of the data points of the angelica which belong to the marine plastic is larger than that of the data points which belong to the non-marine plastic, the class to which the plastic sample to be identified belongs is the marine plastic; and when the data points of the angelica which belongs to the non-marine plastics are larger than the data points which belong to the marine plastics, the class to which the plastic sample to be identified belongs is the non-marine plastics.
If the sample is judged to belong to the marine plastic through the step (3), the plastic sample to be identified is judged to be the marine plastic; otherwise, judging that the plastic sample to be identified is non-marine plastic.
In a preferred embodiment, the process of visual identification of marine plastics is described in conjunction with specific plastic pictures, see fig. 2-4. Fig. 2 is a schematic diagram of a first plastic sample according to an embodiment of the present invention, fig. 3 is a schematic diagram of a second plastic sample according to an embodiment of the present invention, and fig. 4 is a schematic diagram of a third plastic sample according to an embodiment of the present invention.
Visual observation shows that barnacles, coiled worms and bryozoans are attached to one surface of the sample; sample two has ducted insect, barnacle and bryozoan attached; the third sample had the adhesion of the angiozoon and the bryozoan. All the samples are judged as marine plastics.
In a preferred embodiment, the calculation process of the marine plastic similarity, the marine plastic map matching degree and the element content is described with reference to specific plastic pictures, please refer to fig. 5 to 7. Fig. 5 is a schematic diagram of a plastic sample four provided in an embodiment of the present invention, fig. 6 is a schematic diagram of a feature of the plastic sample four provided in the embodiment of the present invention, and fig. 7 is a schematic diagram of a surface adsorbate sampling location of the plastic sample four provided in the embodiment of the present invention.
1) Sampling the plastic sample four shown in the figure 5, then placing the sample under an optical microscope for shooting to obtain the feature structure of the plastic sample four shown in the figure 6, and comparing the feature structure with a preset marine plastic feature database, wherein the feature comparison result is 88%, and the marine plastic similarity of the plastic sample four is 88%;
2) And taking the surface white adsorbent shown in the figure 6 to test the infrared spectrum, and searching with a preset marine plastic infrared spectrum database, wherein the infrared spectrum comparison result is 89.48%, and the marine plastic spectrum matching degree of the plastic sample four is 89.48%.
3) The surface adsorbate was sampled at a position within the frame line in the plastic sample four in fig. 7 and subjected to SEM-EDS elemental analysis to obtain a total amount of surface element contents Na, ca, and Mg of 35.26%, and elemental analysis results of the plastic sample four shown in table 1 were generated.
Table 1: elemental analysis results of Plastic sample No. four
4) The Cl element content of the plastic sample IV was 123.7 ppm by ion chromatography.
The data samples of the plastic sample four before normalization are (88%, 89.48%,35.26% and 123.7 ppm), the data samples of the plastic sample four after normalization are (0.87, 0.89,0.70 and 0.52), the K value after cross-validation optimization is 5, 5 data points which are nearest to the plastic sample four are obtained through calculation of a KNN algorithm, 5 data points which belong to marine plastics are obtained, 0 data point which belongs to non-marine plastics is obtained, and the plastic sample four is judged to be marine plastics.
In a preferred embodiment, the application of high throughput sequencing methods to the identification of plastic samples is described.
The ocean plastic similarity of the plastic sample V is 37%, the ocean plastic map matching degree is 16.48%, the total amount of surface elements Al and Si is 9.6%, and the Cl element content is 62.7 ppm.
The data samples of the plastic sample five before normalization are (37%, 10.49%,9.6%,62.7 ppm), the data samples after normalization are (0.10, 0.07,0.32, 0.18), the optimized K value is 5, and the nearest 5 data points are assigned to 1 marine plastic and 4 non-marine plastics through calculation of a KNN algorithm. Since more data points are attributed to non-marine plastics, re-validation was performed by high throughput sequencing. The statistical results of the five surface microbial populations of the plastic sample accounting for the first 95% by high-throughput sequencing are shown in table 2.
Table 2: statistical result of high-throughput sequencing of surface microorganisms of plastic sample five (first 95%)
Comparing with a marine plastic biological genetic information database, the total content of marine bacteria in the plastic sample five is 79%, calculating by a KNN algorithm, attributing to 5 marine plastic data points and attributing to 0 non-marine plastic data point, and judging that the plastic sample five is marine plastic.
The ocean plastic similarity of the plastic sample six is 41%, the ocean plastic map matching degree is 16.55%, the total amount of surface elements Na, ca and Mg is 5.33%, and the Cl element content is 77.4 ppm.
The data samples of the plastic sample six before normalization are (41%, 16.55%,5.33%,77.4 ppm), the data samples after normalization are (0.16, 0.07,0.11, 0.26), the optimized K value is 5, and the nearest 5 data points are assigned to 0 marine plastics and 5 non-marine plastics by KNN algorithm calculation. The statistical results of the six-surface microbial population of the plastic sample accounting for the first 95% by high-throughput sequencing are shown in Table 3.
Table 3: statistical result of high-throughput sequencing of surface microorganisms of plastic sample six (first 95%)
Comparing with a marine plastic biological genetic information database, the total content of marine bacteria is 2.5%, calculating by a KNN algorithm, attributing to 0 marine plastic data points and 5 non-marine plastic data points, and judging that the plastic sample six is non-marine plastic.
To further explain the marine plastic identification device, please refer to fig. 8, fig. 8 is a schematic structural diagram of a marine plastic identification device according to an embodiment of the present invention, including: an obtaining module 901, a first judging module 902, a calculating module 903 and a second judging module 904;
the acquisition module 901 is used for acquiring corresponding marine plastic similarity, marine plastic map matching degree and element content according to the plastic sample to be identified;
the first determination module 902 is configured to obtain the total content of marine bacteria in the plastic sample to be identified when the plastic sample to be identified is primarily determined to be non-marine plastic by combining a KNN algorithm according to the marine plastic similarity, the marine plastic map matching degree, and the element content;
the calculating module 903 is used for calculating an attribution data point of the total content of the marine bacteria according to the total content of the marine bacteria and a KNN algorithm;
the second determination module 904 is configured to determine that the plastic sample to be identified is marine plastic when the number of data points attributed to marine plastic is greater than the number of data points attributed to non-marine plastic in the attributed data points of the total content of marine bacteria.
In this embodiment, the primary determination that the plastic sample to be identified is non-marine plastic by combining the KNN algorithm according to the marine plastic similarity, the marine plastic map matching degree, and the element content specifically is as follows:
after normalization processing is carried out according to test data in the marine plastic database and the non-marine plastic database, a first characteristic space is established;
performing normalization processing according to the similarity of the marine plastics, the matching degree of the marine plastic map and the element content to serve as a first data sample to be identified;
calculating a Euclidean distance between the first to-be-identified data sample and the first feature space according to a KNN algorithm, and calculating an attribution data point of the first to-be-identified data sample according to the Euclidean distance;
and when the number of the data points attributed to the non-marine plastics in the attributed data points of the first to-be-identified data sample is greater than that of the data points attributed to the marine plastics, determining that the to-be-identified plastic sample is the non-marine plastics for the first time.
In this embodiment, the calculating the attribution data point of the total content of the marine bacteria according to the total content of the marine bacteria and the KNN algorithm specifically includes:
performing normalization processing according to the total content data of the marine bacteria in the marine plastic database and the non-marine plastic database, and establishing a second characteristic space;
taking the total content of the marine bacteria as a second data sample to be identified;
and calculating the Euclidean distance between the second data sample to be identified and the second feature space according to a KNN algorithm, and calculating the attribution data point of the total content of the marine bacteria according to the Euclidean distance.
In a specific embodiment, the obtaining of the corresponding marine plastic similarity according to the plastic sample to be identified specifically includes:
and acquiring the characteristic morphology of the plastic sample to be identified, comparing the characteristic morphology of the plastic sample to be identified with data in a preset marine plastic characteristic morphology database, and calculating and acquiring the corresponding marine plastic similarity of the plastic sample to be identified according to a characteristic morphology comparison result.
In a specific embodiment, the method further comprises:
and acquiring the surface adsorbate infrared spectrum of the plastic sample to be identified, comparing the surface adsorbate infrared spectrum of the plastic sample to be identified with data in a preset marine plastic infrared spectrum database, and calculating according to the infrared spectrum comparison result to obtain the marine plastic spectrum matching degree corresponding to the plastic sample to be identified.
In a specific embodiment, the method further comprises:
and obtaining the surface adsorbate of the plastic sample to be identified, carrying out SEM-EDS test according to the surface adsorbate, and obtaining the surface element content of the plastic sample to be identified according to the test result.
According to the embodiment of the invention, the acquisition module is used for acquiring the corresponding marine plastic similarity, marine plastic map matching degree and element content according to the plastic sample to be identified; when the plastic sample to be identified is primarily judged to be non-marine plastic by combining a KNN algorithm according to the similarity of the marine plastic, the matching degree of the marine plastic spectrum and the element content through a first judgment module, acquiring the total content of marine bacteria of the plastic sample to be identified; calculating an attribution data point of the total content of the marine bacteria according to the total content of the marine bacteria and a KNN algorithm by a calculation module; and when the number of the data points attributed to the marine plastics is greater than that of the data points attributed to the non-marine plastics in the attributed data points of the total content of the marine bacteria through the second judging module, judging that the plastic sample to be identified is the marine plastics.
According to the method, when the plastic sample to be identified is primarily judged to be non-marine plastic by combining a KNN algorithm according to the similarity of the marine plastic, the matching degree of the marine plastic spectrum and the element content, whether the plastic sample to be identified is marine plastic is further identified according to the total content of marine bacteria and the KNN algorithm. According to the method, when the plastic sample to be identified is judged to be non-marine plastic for the first time, the plastic sample to be identified is further judged according to the algorithm and the total content of marine bacteria, so that the problem of misjudgment caused by judgment only according to the similarity of the marine plastic, the matching degree of a marine plastic map and the content of elements is avoided, and the accuracy of identifying the marine plastic is improved; meanwhile, the data of the plastic sample to be identified are automatically identified and classified through the algorithm, so that the problem of low identification precision caused by a manual visual inspection method in the prior art can be solved, and the identification precision and the identification efficiency of the marine plastic are greatly improved.
Furthermore, in the process of obtaining the ocean plastic similarity, the ocean plastic map matching degree, the element content and the total ocean bacteria content of the plastic sample to be identified, the method and the device can detect the plastic sample to be identified step by step according to different characteristics of the plastic sample to be identified, so that the detection efficiency and the result accuracy are improved, and meanwhile, the detection cost can be reduced.
The foregoing is a preferred embodiment of the present invention, and it should be noted that it would be apparent to those skilled in the art that various modifications and enhancements can be made without departing from the principles of the invention, and such modifications and enhancements are also considered to be within the scope of the invention.
Claims (10)
1. A marine plastic identification method, comprising:
acquiring corresponding ocean plastic similarity, ocean plastic map matching degree and element content according to a plastic sample to be identified;
according to the similarity of the marine plastics, the matching degree of the marine plastic spectrum and the element content, when the plastic sample to be identified is primarily judged to be non-marine plastics by combining a KNN algorithm, acquiring the total content of marine bacteria of the plastic sample to be identified;
calculating an attribution data point of the total content of the marine bacteria according to the total content of the marine bacteria and a KNN algorithm;
and when the number of the data points attributed to the marine plastics is greater than that of the data points attributed to the non-marine plastics in the attributed data points of the total content of the marine bacteria, determining that the plastic sample to be identified is the marine plastics.
2. The marine plastic identification method according to claim 1, wherein the to-be-identified plastic sample is primarily determined to be non-marine plastic by combining a KNN algorithm according to the marine plastic similarity, the marine plastic map matching degree and the element content, and specifically comprises the following steps:
after normalization processing is carried out according to test data in the marine plastic database and the non-marine plastic database, a first characteristic space is established;
performing normalization processing according to the ocean plastic similarity, the ocean plastic map matching degree and the element content to serve as a first data sample to be identified;
calculating a Euclidean distance between the first to-be-identified data sample and the first feature space according to a KNN algorithm, and calculating an attribution data point of the first to-be-identified data sample according to the Euclidean distance;
and when the number of the data points attributed to the non-marine plastics in the attributed data points of the first to-be-identified data sample is greater than that of the data points attributed to the marine plastics, determining that the to-be-identified plastic sample is the non-marine plastics for the first time.
3. The method for identifying marine plastics according to claim 2, wherein the attribution data points of the total content of marine bacteria are calculated according to the total content of marine bacteria and a KNN algorithm, and specifically are as follows:
performing normalization processing according to the total content data of the marine bacteria in the marine plastic database and the non-marine plastic database, and establishing a second characteristic space;
taking the total content of the marine bacteria as a second data sample to be identified;
and calculating the Euclidean distance between the second data sample to be identified and the second feature space according to a KNN algorithm, and calculating the attribution data point of the total content of the marine bacteria according to the Euclidean distance.
4. The marine plastic identification method according to claim 3, wherein the obtaining of the corresponding marine plastic similarity according to the plastic sample to be identified specifically comprises:
and acquiring the characteristic morphology of the plastic sample to be identified, comparing the characteristic morphology of the plastic sample to be identified with data in a preset marine plastic characteristic morphology database, and calculating and acquiring the corresponding marine plastic similarity of the plastic sample to be identified according to a characteristic morphology comparison result.
5. The marine plastic identification method of claim 4, further comprising:
and acquiring the surface adsorbate infrared spectrum of the plastic sample to be identified, comparing the surface adsorbate infrared spectrum of the plastic sample to be identified with data in a preset marine plastic infrared spectrum database, and calculating to obtain the marine plastic spectrum matching degree corresponding to the plastic sample to be identified according to the infrared spectrum comparison result.
6. The marine plastic identification method of claim 5, further comprising:
and obtaining the surface adsorbate of the plastic sample to be identified, carrying out SEM-EDS test according to the surface adsorbate, and obtaining the surface element content of the plastic sample to be identified according to the test result.
7. A marine plastic identification device, comprising: the device comprises an acquisition module, a first judgment module, a calculation module and a second judgment module;
the acquisition module is used for acquiring corresponding marine plastic similarity, marine plastic map matching degree and element content according to a plastic sample to be identified;
the first judging module is used for acquiring the total content of marine bacteria of the plastic sample to be identified when the plastic sample to be identified is primarily judged to be non-marine plastic by combining a KNN algorithm according to the similarity of the marine plastic, the matching degree of the marine plastic spectrum and the element content;
the calculation module is used for calculating an attribution data point of the total content of the marine bacteria according to the total content of the marine bacteria and a KNN algorithm;
and the second determination module is used for determining that the plastic sample to be identified is the marine plastic when the number of the data points which belong to the marine plastic is greater than the number of the data points which belong to the non-marine plastic in the attribution data points of the total content of the marine bacteria.
8. The marine plastic identification device according to claim 7, wherein the to-be-identified plastic sample is primarily determined to be non-marine plastic by combining a KNN algorithm according to the marine plastic similarity, the marine plastic map matching degree and the element content, and specifically:
after normalization processing is carried out according to test data in the marine plastic database and the non-marine plastic database, a first characteristic space is established;
performing normalization processing according to the ocean plastic similarity, the ocean plastic map matching degree and the element content to serve as a first data sample to be identified;
calculating a Euclidean distance between the first to-be-identified data sample and the first feature space according to a KNN algorithm, and calculating an attribution data point of the first to-be-identified data sample according to the Euclidean distance;
and when the number of the data points attributed to the non-marine plastics in the attributed data points of the first to-be-identified data sample is greater than that of the data points attributed to the marine plastics, determining that the to-be-identified plastic sample is the non-marine plastics for the first time.
9. The marine plastic identification device according to claim 8, wherein said calculating said attributed data points of said total content of marine bacteria according to said total content of marine bacteria and KNN algorithm is specifically:
performing normalization processing according to the total content data of the marine bacteria in the marine plastic database and the non-marine plastic database, and establishing a second characteristic space;
taking the total content of the marine bacteria as a second data sample to be identified;
and calculating the Euclidean distance between the second data sample to be identified and the second feature space according to a KNN algorithm, and calculating the attribution data point of the total content of the marine bacteria according to the Euclidean distance.
10. The marine plastic identification device according to claim 9, wherein the obtaining of the corresponding marine plastic similarity according to the plastic sample to be identified is specifically:
and acquiring the characteristic features of the plastic sample to be identified, comparing the characteristic features of the plastic sample to be identified with data in a preset marine plastic characteristic feature database, calculating the corresponding marine plastic similarity of the plastic sample to be identified according to the characteristic feature comparison result, and acquiring the corresponding marine plastic similarity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211645479.5A CN115639354B (en) | 2022-12-21 | 2022-12-21 | Marine plastic identification method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211645479.5A CN115639354B (en) | 2022-12-21 | 2022-12-21 | Marine plastic identification method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115639354A CN115639354A (en) | 2023-01-24 |
| CN115639354B true CN115639354B (en) | 2023-03-21 |
Family
ID=84949555
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211645479.5A Active CN115639354B (en) | 2022-12-21 | 2022-12-21 | Marine plastic identification method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115639354B (en) |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20230314314A1 (en) * | 2020-05-11 | 2023-10-05 | Woods Hole Oceanographic Institution | Optical system and method to identify plastic |
| KR102432313B1 (en) * | 2021-01-08 | 2022-08-11 | 강원대학교산학협력단 | Detection method for microplastic |
| CN113535977B (en) * | 2021-07-13 | 2023-12-19 | 泰康保险集团股份有限公司 | Knowledge graph fusion method, device and equipment |
| CN114359238A (en) * | 2022-01-07 | 2022-04-15 | 自然资源部第三海洋研究所 | Method, device and storage medium for identifying micro-plastic |
| CN114674803A (en) * | 2022-02-16 | 2022-06-28 | 北京化工大学 | Quantitative analysis method for components of marine environment micro-plastic mixture |
| CN114998664A (en) * | 2022-07-18 | 2022-09-02 | 中国科学院烟台海岸带研究所 | Rapid detection method and device for micro-plastic in seawater by multiple optical platforms |
| CN115187870B (en) * | 2022-09-13 | 2023-01-03 | 浙江蓝景科技有限公司杭州分公司 | Marine plastic waste material identification method and system, electronic equipment and storage medium |
-
2022
- 2022-12-21 CN CN202211645479.5A patent/CN115639354B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN115639354A (en) | 2023-01-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xiong et al. | Zooplankton biodiversity monitoring in polluted freshwater ecosystems: A technical review | |
| Gorsky et al. | Digital zooplankton image analysis using the ZooScan integrated system | |
| Pennekamp et al. | Implementing image analysis in laboratory‐based experimental systems for ecology and evolution: a hands‐on guide | |
| EP2414536B1 (en) | Fluorescence cued raman identification of viable organisms | |
| Soltis et al. | Plants meet machines: Prospects in machine learning for plant biology | |
| Deglint et al. | Investigating the automatic classification of algae using the spectral and morphological characteristics via deep residual learning | |
| CN115239209B (en) | Method, equipment and system for screening heavy metal-containing humus soil in refuse landfill | |
| CN114399763B (en) | Single-sample and small-sample micro-body paleobiological fossil image identification method and system | |
| CN113295835B (en) | Remote sensing-based micro plastic polluted water area spatial distribution identification method | |
| Gislason et al. | Comparison between automated analysis of zooplankton using ZooImage and traditional methodology | |
| CN113658640B (en) | Fresh water ecological system health evaluation method | |
| CN114359238A (en) | Method, device and storage medium for identifying micro-plastic | |
| CN115639354B (en) | Marine plastic identification method and device | |
| CN113155753A (en) | Farmland surface soil micro-plastic detection method and system | |
| CN117172430A (en) | Deep learning-based water body environment assessment and prediction method and system | |
| Culverhouse et al. | An instrument for rapid mesozooplankton monitoring at ocean basin scale | |
| CN114925759A (en) | Feature analysis method for Ether fishing behavior account | |
| Ershova et al. | Mock samples resolve biases in diversity estimates and quantitative interpretation of zooplankton metabarcoding data | |
| Li | Bivariate and trivariate analysis in flow cytometry: Phytoplankton size and fluorescence | |
| CN111833297A (en) | Disease association method of marrow cell morphology automatic detection system | |
| CN113588588B (en) | Airborne micro-plastic rapid detection method based on ground physical spectrometer | |
| CN117171675B (en) | Water environment microorganism detection method, system and medium based on multi-source data | |
| Li et al. | A Deep Learning based Method for Microscopic Object Localization and Classification | |
| Mary et al. | Freshwater microalage image identification and classification based on machine learning technique | |
| CN114863426B (en) | Micro target detection method for coupling target feature attention and pyramid |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |