A device and method for rapid detection of microplastics based on backscattering polarimetry

Abstract

A kind of microplastic rapid detection device and method based on backward polarization imaging, comprising: polarization state generator, generating polarized light irradiation microplastic in water body;Polarization state receiver, receive the polarization information of backscattering polarized light, obtain the polarization image containing polarization parameter;Polarization image detection and identification device, split the polarization image into multiple polarization pixel points, each polarization pixel point corresponds a one-dimensional vector containing multiple elements, each element value in one-dimensional vector represents the polarization value of corresponding polarization pixel point under multiple different polarization states, so as to capture the complete polarization characteristics of each polarization pixel point;Through one-dimensional data classification model 1D-ResNet, each polarization pixel point is classified, and the classification label of each polarization pixel point is obtained;The classification result is mapped back to image space, to realize the rapid detection and visual identification of microplastic in water body.The present application can detect microplastic particles in water body more quickly and accurately compared with existing detection techniques.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of optical measurement technology, in particular to a microplastic rapid detection device and method based on backward polarization imaging. BACKGROUND

[0002] Microplastics are generally defined as plastic particles with a particle size of less than 5 millimeters, mainly from the fragmentation and degradation of plastic products. In recent years, microplastics have become one of the major global environmental problems as a new type of man-made pollutant, and their distribution in water bodies has received much attention from the scientific community, society and the public. Due to human emissions and water transport, the content of microplastics in water environments is increasing and is difficult to collect and clean. Microplastics are easily ingested by aquatic organisms, causing sustained negative effects on biological health. Since they cannot be excreted from the body through metabolism, microplastics accumulate in the food chain, causing serious damage to aquatic ecosystems. At the same time, microplastics can also be ingested by the human body through the food chain and drinking water, which may cause a series of health problems and pose a potential threat to human life. Therefore, it is urgent to develop detection methods and devices for microplastics to achieve rapid and accurate statistics of the abundance and distribution of microplastics, strengthen the monitoring and management of microplastics in water environments, and provide strong data support for real-time monitoring of microplastic pollution, identification of emission sources, and accurate water treatment.

[0003] Microplastic samples in water bodies are generally obtained by sampling and pretreatment. The detection methods for microplastics after pretreatment include visual method, spectroscopic method, thermal analysis method, etc. The visual method uses naked eye or microscope to observe the particulate matter, count and record its color, shape and other information. The visual method is simple to operate, but cannot judge the composition of microplastics and has low accuracy. In addition, since the light source of the traditional projection microscope cannot penetrate thick samples, the visual method has higher requirements for samples and often requires tedious pretreatment operations. The spectroscopic method detects and analyzes the sample by Fourier transform infrared spectroscopy, Raman spectroscopy and other technologies, which has the characteristics of non-invasive and high accuracy. Spectroscopic method takes a long time, so it is usually combined with visual method to select suspected microplastic particulate samples by visual method and then analyze them by spectroscopic method. This can improve the detection accuracy of microplastics, but the use of multiple instruments not only tests the professional skills of the operator, but also increases the detection cost and time consumption.

[0004] CN116840120A discloses a water body suspended particle flow polarization imaging and classification device, comprising a flow liquid inlet system, a collection module and a processing module, the collection module comprising a detection cavity, a flash lamp, a polarizer and an image-polarization detection module; the flow liquid inlet system is in communication with the detection cavity, for injecting a sample to be measured into the detection cavity; the flash lamp is used to emit a non-polarized light beam with uniform intensity; the polarizer is used to adjust the non-polarized light beam emitted by the flash lamp to a predetermined polarization state and keep it constant during sampling; the image-polarization detection module is used to collect original images of the sample to be measured in the detection cavity at a fixed frequency; the processing module is used to process the original images to obtain Stokes images, and obtain the polarization information and morphological information of the suspended particles in the sample to be measured according to the Stokes images. However, this device has limitations and deficiencies in realizing rapid and accurate detection and identification of microplastic particles. When extracting and classifying the feature information of microplastics from the original polarization image, it is difficult to efficiently and accurately automatically analyze and identify the unique optical properties of microplastic particles, lacks the ability to quickly and accurately identify and classify microplastic particles in the polarization image, and requires relatively complex image preprocessing and threshold segmentation steps.

[0005] Therefore, there is still an urgent need to develop new solutions to more accurately and quickly detect microplastic particles.

[0006] It should be noted that the information disclosed in the above background section is only for understanding the background of the present application, and therefore can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0007] The main purpose of the present application is to solve the problems existing in the above background technology, and to provide a microplastic rapid detection device and method based on backward polarization imaging.

[0008] To achieve the above-mentioned purpose, the present application adopts the following technical solutions:

[0009] A microplastic rapid detection device based on backward polarization imaging, comprising:

[0010] A polarization state generator for generating polarized light of a specific polarization state to irradiate a microplastic sample in water;

[0011] A polarization state receiver for receiving polarization information of backscattered polarized light to obtain a polarization image containing polarization parameters;

[0012] The polarization image detection and recognition device is configured to: split a polarization image into a plurality of polarization pixel points, each polarization pixel point corresponding to a one-dimensional vector containing a plurality of elements, and each element value in the one-dimensional vector representing a polarization value of the corresponding polarization pixel point under a plurality of different polarization states, so as to capture the complete polarization characteristics of each polarization pixel point; classify each polarization pixel point through a one-dimensional data classification model 1D-ResNet to obtain a classification label of each polarization pixel point; and remap the classification result back to the image space to realize rapid detection and visual recognition of each microplastic sample.

[0013] A microplastic rapid detection method using the microplastic rapid detection device, comprising:

[0014] The polarization state generator generates polarized light of a specific polarization state and irradiates a microplastic sample in a water body;

[0015] The polarization state receiver receives polarization information of the backscattered polarized light to obtain a polarization image containing polarization parameters;

[0016] The polarization image detection and recognition device splits the polarization image into a plurality of polarization pixel points, each polarization pixel point corresponding to a one-dimensional vector containing a plurality of elements, and each element value in the one-dimensional vector representing a polarization value of the corresponding polarization pixel point under a plurality of different polarization states, so as to capture the complete polarization characteristics of each polarization pixel point; classifies each polarization pixel point through a one-dimensional data classification model 1D-ResNet to obtain a classification label of each polarization pixel point; and remaps the classification result back to the image space to realize rapid detection and visual recognition of each microplastic sample.

[0017] The present application has the following beneficial effects:

[0018] The present application provides an innovative microplastic rapid detection device and method based on back-polarization imaging. In the present application, the polarization state generator generates polarized light of a specific polarization state and irradiates microplastics in a water body, the polarization state receiver receives backscattered polarized light to obtain a polarization image containing polarization parameters, the polarization image is split into a plurality of polarization pixel points to form a one-dimensional vector, and each pixel point is classified through a one-dimensional data classification model 1D-ResNet, thereby realizing pixel-level recognition and classification of microplastic particles. The solution of the present application not only avoids complex image preprocessing and threshold segmentation, but also realizes rapid detection and visual recognition by remapping the classification result to the image space, significantly improving the detection efficiency and accuracy. In addition, the present application allows measurement of microplastics of different morphologies and thicknesses without the need for slicing, staining and other pretreatments, further simplifying the detection process, making it a fast and accurate detection technology with practical application value for environmental microplastic pollution problems.

[0019] Other benefits of embodiments of the present application will be described further below. BRIEF DESCRIPTION OF DRAWINGS

[0020] Figure 1 A schematic diagram of an optical path structure of an embodiment of the present application.

[0021] Figure 2 A flowchart of a detection algorithm of an embodiment of the present application.

[0022] Figure 3 A schematic diagram of an optical path structure of a preferred embodiment of the present application.

[0023] Figure 4 A flowchart of an implementation method of microplastic pixel classification of an embodiment of the present application. DETAILED DESCRIPTION

[0024] The embodiments of the present application will be described in detail below. It should be emphasized that the following description is only exemplary and is not intended to limit the scope of the present application and its applications.

[0025] It should be noted that when an element is referred to as being “fixed” or “set” on another element, it can be directly on the other element or indirectly on the other element. When an element is referred to as being “connected” to another element, it can be directly connected to the other element or indirectly connected to the other element. In addition, the connection can be for fixing or for coupling or communicating.

[0026] It should be understood that the terms “length”, “width”, “upper”, “lower”, “front”, “back”, “left”, “right”, “vertical”, “horizontal”, “top”, “bottom”, “inner”, “outer” and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the embodiments of the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as limiting the present application.

[0027] In addition, the terms “first”, “second” are only for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the technical features indicated. Therefore, the features defined with “first”, “second” can explicitly or implicitly include one or more of the features. In the description of the embodiments of the present application, the meaning of “a plurality of” is two or more, unless otherwise specifically limited.

[0028] Reference Figures 1 to 4The embodiment of the present application provides a kind of based on backward polarization imaging microplastic rapid detection device, comprising: polarization state generator, for generating the polarized light of specific polarization state, irradiation microplastic sample in water body;Polarization state receiver, for receiving the polarization information of backscattering polarized light, to obtain the polarization image containing polarization parameter, the polarization parameter can include Stokes vector, Mueller matrix and / or its derived polarization quantity;Polarization image detection identification device is configured to: utilize pixel split technique, polarization image is decomposed into multiple polarization pixel points, each polarization pixel point corresponds one one-dimensional vector containing multiple elements, the length of vector is determined by the depth of measured polarization information (i.e. the number of polarization state), each element value in the one-dimensional vector represents the polarization value of corresponding polarization pixel point under multiple different polarization states, so as to capture the complete polarization characteristics of each polarization pixel point;Using one-dimensional data classification model 1D-ResNet, each polarization pixel point is classified, and the classification label of each polarization pixel point is obtained;Classification result is mapped back to image space, and the rapid detection and visual identification of each microplastic sample are realized.

[0029] The present application adopts backward polarization imaging device to obtain polarization image in visual field, and the pixel level recognition and classification of microplastic particles are realized by innovative rapid detection method, so that the rapid detection and visual identification of millimeter and sub-millimeter microplastic samples are realized.Especially, unlike traditional image classification scheme, the present application proposes the whole disassembly idea by innovative pixel classification method, for the obvious morphological difference of microplastic particles, wide particle size range, and the composition of particulate matter in real scene is complex, the traditional detection method of image preprocessing and image segmentation is difficult to realize fast and accurate measurement, the present application innovatively proposes the whole disassembly idea, splits polarization image into several polarization pixel points containing polarization information, identifies and classifies using the polarization information contained in each pixel point, and maps the classification result back to image space, to realize the detection and visual identification of microplastic quickly and accurately.

[0030] As shown in Figure 3 In the preferred embodiment, the polarization state generator includes collimating lens, linear polarizer and quarter-wave plate, for forming incident light of different polarization states.

[0031] As shown in Figure 3As shown, in the preferred embodiment, the polarization state receiver comprises a collimating lens, a non-polarized beam splitter, a quarter wave plate and at least two cameras for acquiring Stokes images and Mueller images. Preferably, the polarization state receiver is configured such that the backscattered light from the sample is collected by the collimating lens along the sample surface normal direction, again forming a parallel light beam, which is then split by the non-polarized beam splitter into two beams of the same polarization state in a 50:50 ratio, wherein the transmitted light is collected by one camera after passing through the fixed quarter wave plate, and the reflected light is collected by the other camera, and the polarization images taken by the two cameras are processed to obtain the Stokes images of the microplastic sample. Preferably, the Stokes images under different polarization states of incidence can be obtained by rotating the angle of the quarter wave plate at least four times, and the corresponding Mueller images are calculated.

[0032] As shown, in some embodiments, the angle between the polarization state generator and the polarization state receiver can be dynamically adjusted to optimize the measurement. Figure 1

[0033] In the preferred embodiment, the polarization image detection and recognition device maps the obtained classification results to the original image space by color coding; wherein the polarization image detection and recognition device is configured with a pre-defined color mapping table for corresponding different classification labels with specific colors to achieve intuitive differentiation and visual recognition of the background and different plastic types.

[0034] The embodiments of the present application also provide a microplastic rapid detection method based on backscattering polarization imaging, which uses the detection device of any of the preceding embodiments to achieve rapid detection and visual recognition of each microplastic sample.

[0035] The specific embodiments of the present application are further described below.

[0036] Figure 1 The optical path structure of the microplastic rapid detection device is shown in the figure. The light source passes through the polarization state generator to generate a light beam of a specific polarization state, and the incident light beam is obliquely incident on the microplastic sample to form a uniform illumination area. Then, the scattered light of the sample is acquired by the polarization state receiver to obtain image information of the sample under measurement; the image information is transmitted to the signal processing unit, and the polarization image of the microplastic can be obtained by image analysis. Using image processing, feature parameter extraction, classification recognition and other algorithms, the number, particle size, morphology, color and composition of different microplastic samples in the field of view can be obtained, and rapid detection of microplastics can be achieved.

[0037] ​(1) Rapid measurement of microplastic polarization image. Microplastic particles have obvious morphological differences. Traditional projection imaging cannot penetrate thicker microplastics, and must be pre-processed by slicing and staining. Therefore, the present application designs a device based on backscattering polarization imaging, which can obtain the backscattering polarization image of microplastics without complicated pretreatment operations. As shown in Figure 1 , the polarization state generator can generate polarized light with a specific polarization state to realize continuous and stable irradiation of the microplastic sample; the polarization state receiver can receive the polarization information of the backscattering polarized light, and obtain the Stokes vector, Mueller matrix or derived polarization quantity of the image through the signal processing unit to meet the accurate identification and detection of microplastics.

[0038] Polarization is one of the inherent properties of light. Polarization measurement has the characteristics of non-contact, non-destructive and high resolution. The polarization state of light can be represented by a 4x1 Stokes (S) vector,

[0039]

[0040] where I represents the light intensity, Q, U and V are polarization components. Q represents the intensity difference between the horizontal component IH and the vertical component I V of the light, U represents the intensity difference between the 45° linear component I 45 and the 135° linear component I 135 of the light, and V represents the intensity difference between the right circular component I R and the left circular component I L of the light. Since the polarization state can be modulated by polarizing plates and wave plates without changing the propagation direction of light, the polarization optical device can be compatible with the traditional non-polarization measurement optical path. When the incident polarized light irradiates the object, the polarization state S in of the incident light becomes the polarization state S out of the scattered light through linear transformation, which is represented as follows:

[0041] S out = M x S in #(2)

[0042] where M is a 4x4 Mueller matrix:

[0043]

[0044] The 16 elements of Mueller matrix contain all the polarization properties of suspended particles. Polarized scattered light is very sensitive to the optical properties of the object (including refractive index, absorption, birefringence, dichroism, etc.), morphology (particle size, shape, orientation, arrangement, order, etc.) and microstructure (surface topography, internal substructure), especially to the super-optical resolution structure below the subwavelength scale, which can express the rich microstructure information of particles. Using Mueller matrix which can completely describe the polarization optical properties of microplastics can achieve detailed classification of microplastic composition, and can quantitatively characterize the microstructure characteristics of microplastics.

[0045] During the measurement process, due to the strong surface reflection of the microplastic sample when the light beam irradiates it, overexposure occurs, the image quality decreases, and the measurement of the microplastic polarization image is seriously disturbed, therefore, by controlling the angle between the polarization state generator and the polarization state receiver, the distortion of the imaging result is not caused. The back oblique incidence angle θ of the device can also be dynamically adjusted to achieve optimal measurement. In addition, the measurement calibration of the instrument can eliminate the errors of the external environment and the instrument itself to ensure the consistency of the measurement.

[0046] (2) Rapid detection of microplastic polarization image. Unlike traditional image classification methods, the present application adopts an innovative pixel classification method. The morphology of microplastic particles is obviously different, the particle size range is wide, and the composition of particles in real scenes is complex, so it is difficult to achieve rapid and accurate measurement by using traditional detection methods such as image preprocessing and image segmentation. The innovative method of the present application adopts the idea of overall disassembly, splits the polarization image into a plurality of polarization pixel points containing polarization information, and uses the polarization information contained in each pixel point for identification and classification.

[0047] The rapid detection algorithm flow of the present application is shown in Figure 2 The polarization parameters (Stokes vector, Mueller matrix or derived polarization quantities) measured by the device are first input, and the image is split into pixel points containing polarization information, i.e. a plurality of one-dimensional vectors. The vector length is determined by the depth of the measured polarization information. Then, a one-dimensional classification model is used to classify these pixel points to obtain the type of each pixel point. Finally, the classification results are mapped back to the image space to visually identify the types of microplastics in different regions of the image. This method decouples microplastic classification and image recognition, thereby reducing the interference of microplastic morphology on identification and classification. In addition, this method directly inputs the original polarization image measured by the device, without the need for image preprocessing and image segmentation operations required by traditional image recognition methods, greatly improving the detection efficiency and achieving rapid and accurate detection of microplastics.

[0048] Examples

[0049] (1) Rapid measurement of microplastic polarization image. Figure 3An implementation of the device. The polarization state generator is composed of a collimating lens, a linear polarizer and a quarter-wave plate. The LED light source forms a parallel light beam through the collimating lens 1, and then passes through the linear polarizer and the rotatable quarter-wave plate to obtain incident light of different polarization states. The illumination beam is obliquely incident on the sample, forming an almost uniform illumination area with a diameter of 5 cm on the sample. The polarization state receiver is composed of a collimating lens, a non-polarized beam prism, a quarter-wave plate and two cameras. The backscattered light from the sample is collected by the collimating lens 2 along the normal direction of the sample surface, forming a parallel light beam again. Then, the non-polarized beam prism divides the light beam into two beams with the same polarization state at a ratio of 50:50. Among them, the transmitted light is collected by camera 1 after passing through the fixed quarter-wave plate, and the reflected light is directly collected by camera 2. The polarization images taken by the two cameras can be obtained after image processing by the image processing unit. To obtain the Mueller image, the angle of the quarter-wave plate 1 needs to be rotated at least four times to obtain the Stokes image under different polarization states, and then the Mueller image under the field of view is calculated by formula 2. In practical application, the polarization image of microplastic can be obtained flexibly according to the requirements of measurement speed and measurement accuracy.

[0050] In order to avoid the influence of the reflection of the sample surface on the measurement, it is necessary to control the angle θ between the polarization state generator and the polarization state receiver. Due to the different reflection abilities of different microplastic sample surfaces, the angle θ can be changed during measurement to realize the optimal measurement of the polarization image of microplastic. In order to reduce the instrument error and the interference of the external environment, the method measures the air as a standard sample before measuring the microplastic sample to calibrate the system.

[0051] (2) Rapid detection of microplastic polarization image. Figure 4 An implementation of the detection method, which combines pixel splitting and one-dimensional data classification model (1D-ResNet) to realize rapid and accurate detection of microplastic samples. Specifically, first, the device obtains the images of 16 matrix elements by Mueller matrix imaging of the sample, and the size of each image is n x m. The 16 x n x m dimensional image information is input into the algorithm, and each pixel point in the image is processed by splitting to obtain a one-dimensional vector of (n x m) elements, and the length of the vector of each pixel point is 16. After extracting all the polarization pixels in the image, the 1D-ResNet model is used for pixel classification, and the specific label of each pixel point in t categories is obtained. After the classification of all pixel points is completed, the original n x m image space is remapped to obtain the visual classification result, and the background and plastic and different plastic types are visualized and identified according to the color of different labels, realizing the rapid detection of each microplastic sample.

[0052] In summary, the present application provides a rapid detection device and method for backward polarization imaging microplastics. In order to better meet the detection needs of microplastic samples of different morphologies and thicknesses, the device measures the polarization image of the sample, the incident light is obliquely incident on the sample to form a uniform illumination area, and the sample can be measured without pretreatment methods such as slicing and staining, greatly improving the detection efficiency. In particular, the method of the present application does not need to be pretreated and threshold segmented, but first splits the polarization image into a plurality of polarization pixel points, classifies the pixel points containing polarization information, and then remaps the classification results to the image space, thereby realizing rapid detection and visual recognition of microplastics. The present application splits the polarization image into pixel points containing polarization information, i.e. a plurality of one-dimensional vectors. The vector length is determined by the depth of the measured polarization information. Then, a one-dimensional classification model is used to classify these pixel points to obtain the type of each pixel point. Finally, the classification results are remapped to the image space to visually identify the types of microplastics in different regions of the image. The present application decouples microplastic classification and image recognition, thereby reducing the interference of microplastic morphology on recognition and classification. In addition, the present application directly inputs the original polarization image measured by the device, without the need for image preprocessing and image segmentation operations required by traditional image recognition methods, greatly improving the detection efficiency and realizing rapid and accurate detection of microplastics.

[0053] The above is a further detailed description of the present application in combination with specific / preferred embodiments, and cannot be considered as limiting the specific implementation of the present application to these descriptions. For those of ordinary skill in the art to which the present application belongs, without departing from the concept of the present application, they can make several alternatives or modifications to the described embodiments, and these alternatives or modifications shall be considered as falling within the protection scope of the present application. In the description of the present specification, the description of the terms "an embodiment", "some embodiments", "preferred embodiment", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in combination with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. Those skilled in the art can combine and combine the different embodiments or features of the embodiments or examples described in the present specification without contradiction. Although embodiments of the present application and their advantages have been described in detail, it should be understood that various changes, substitutions and modifications can be made herein without departing from the scope of the patent application.

Claims

1. A device for rapid detection of microplastics based on backscattering polarimetry, characterized in that, The application relates to a polarization state generator for generating polarized light of specific polarization states to irradiate microplastics in a water body; a polarization state receiver for receiving polarization information of backscattered polarized light to obtain a polarization image containing polarization parameters, wherein the polarization parameters of the polarization image include a Stokes vector, a Mueller matrix and / or derived polarization quantities thereof; and a polarization image detection and recognition device configured to: split the polarization image into a plurality of polarization pixels, each of which corresponds to a one-dimensional vector containing a plurality of elements, wherein each element value in the one-dimensional vector represents a polarization value of the corresponding polarization pixel under a plurality of different polarization states, so as to capture the complete polarization characteristics of each polarization pixel; classify each polarization pixel through a one-dimensional data classification model 1D-ResNet to obtain a classification label of each polarization pixel; and remap the classification result back to the image space to realize rapid detection and visual recognition of the microplastics. The polarization state generator comprises a collimating lens, a linear polarizer and a rotatable quarter-wave plate for forming incident light of different polarization states. The polarization state receiver comprises a collimating lens, a non-polarized beam prism, a fixed quarter-wave plate and at least two cameras for obtaining Stokes images and Mueller images. The polarization state receiver is configured to collect backscattered light from the sample along the normal direction of the sample surface by the collimating lens to form a parallel light beam again, and then split the light beam into two light beams with the same polarization state at a ratio of 50:50 by the non-polarized beam prism, wherein the transmitted light is collected by one camera after passing through the fixed quarter-wave plate, and the reflected light is collected by another camera, and the polarization images captured by the two cameras are processed to obtain the Stokes images of the microplastic sample.

2. The rapid microplastics detection device according to claim 1, wherein, The Stokes images under different polarization states are obtained by rotating the angle of the rotatable quarter-wave plate at least four times, and the corresponding Mueller images are calculated.

3. The rapid microplastics detection device according to claim 2, wherein, The angle between the polarization state generator and the polarization state receiver is dynamically adjusted to realize optimal measurement.

4. The rapid microplastics detection device according to claim 3, wherein, The polarization image detection and recognition device maps the obtained classification result to the original image space through color coding; wherein the polarization image detection and recognition device is configured with a predefined color mapping table for corresponding different classification labels with specific colors to realize intuitive differentiation and visual recognition of the background and different plastic types.

5. The rapid microplastics detection device according to claim 4, wherein, The application relates to a polarization state generator for generating polarized light of specific polarization states to irradiate microplastics in a water body; a polarization state receiver for receiving polarization information of backscattered polarized light to obtain a polarization image containing polarization parameters, wherein the polarization parameters of the polarization image include a Stokes vector, a Mueller matrix and / or derived polarization quantities thereof; and a polarization image detection and recognition device configured to: split the polarization image into a plurality of polarization pixels, each of which corresponds to a one-dimensional vector containing a plurality of elements, wherein each element value in the one-dimensional vector represents a polarization value of the corresponding polarization pixel under a plurality of different polarization states, so as to capture the complete polarization characteristics of each polarization pixel; classify each polarization pixel through a one-dimensional data classification model 1D-ResNet to obtain a classification label of each polarization pixel; and remap the classification result back to the image space to realize rapid detection and visual recognition of the microplastics.

6. The rapid microplastics detection device according to any one of claims 1 to 5, wherein, ​ 7. The rapid microplastics detection device according to any one of claims 1 to 5, wherein, ​ 8. A method for rapid detection of microplastics using the rapid detection device for microplastics according to any one of claims 1 to 7, characterized in that, ​ ​ ​ The polarization image detection and recognition device splits the polarization image into a plurality of polarization pixel points, each polarization pixel point corresponds to a one-dimensional vector containing a plurality of elements, and each element value in the one-dimensional vector represents the polarization value of the corresponding polarization pixel point in a plurality of different polarization states, thereby capturing the complete polarization characteristics of each polarization pixel point; through a one-dimensional data classification model 1D-ResNet, each polarization pixel point is classified to obtain the classification label of each polarization pixel point; the classification result is mapped back to the image space to realize the rapid detection and visual recognition of microplastics.

9. The method for rapid detection of microplastics according to claim 8, wherein, Also included is calibrating the system by measuring air as a standard sample before measurement.

Images