A method for quantifying polystyrene microplastic particles

Abstract

This invention relates to the field of quantitative identification technology for plastic particles, and discloses a method for quantitatively identifying polystyrene microplastic particles. The method includes real-time acquisition of images of particles during transmission, identification of the dominant target color, automatic switching of the imaging background to the color with the highest contrast based on the target color to enhance the visual distinguishability of dissimilar particles, simultaneous acquisition of the two-dimensional position information and three-dimensional height information of the particles, construction of a spatial state vector for each particle, identification of dissimilar particles (not the target color) based on color classification, assignment of unique identifiers to each particle, and tracking of the motion trajectory of the dissimilar particles using multiple consecutive frames of images. This invention achieves micron-level three-dimensional positioning in high-speed online scenarios, providing crucial depth information for subsequent trajectory prediction and layered counting, fundamentally solving the problem that traditional two-dimensional imaging cannot distinguish overlapping particles in upper and lower layers, thus laying the foundation for high-precision measurement.

Description

Technical Field

[0001] This invention relates to the field of quantitative identification technology for plastic particles, specifically a method for quantitatively identifying polystyrene microplastic particles. Background Technology

[0002] Plastic granules are plastic raw materials that exist in granular form. They are classified into three categories according to their properties: general plastics, engineering plastics, and specialty plastics. General plastics include common types such as polypropylene and polyethylene. Microplastics specifically refer to artificial polymer particles with a particle size of less than 5 mm, including industrial product additives and plastic degradation products. When processing polystyrene microplastic particles, it is necessary to periodically check the processing efficiency and state of the plastic particles. Therefore, it is necessary to quantitatively identify polystyrene microplastic particles on the production line.

[0003] Existing quantitative identification methods for polystyrene microplastic particles can only detect the approximate quantity on the production line and cannot make accurate judgments. Furthermore, when a large number of mixed-color particles are mixed in, it is impossible to distinguish the quantity of each particle. The quantity of plastic particles can only be determined after the color sorter has screened them. However, the color sorter cannot dynamically detect polystyrene microplastic particles on the production line in real time. Summary of the Invention

[0004] This invention provides a method for quantifying polystyrene microplastic particles, which has the beneficial effects of dynamic high-contrast screening and high-precision capture and quantitative calculation.

[0005] This invention provides the following technical solution: a method for quantitatively measuring polystyrene microplastic particles, comprising:

[0006] Real-time acquisition of images of particles during transmission, and identification of the dominant target color;

[0007] The imaging background is automatically switched to the color with the highest contrast based on the target color to enhance the visual distinguishability of dissimilar particles.

[0008] Simultaneously acquire the two-dimensional position information and three-dimensional height information of the particles to construct the spatial state vector of each particle;

[0009] Based on color classification, non-target color particles are identified and assigned a unique identifier.

[0010] The motion trajectory of the heterochromatic particles is tracked using multiple consecutive images, and their parabolic motion model is fitted to predict the impact point.

[0011] The vibration parameters of the vibration platform and the direction of the edge airflow are dynamically adjusted according to the predicted landing point to directionally throw the discolored particles to the rejection area.

[0012] The particles are assigned to multiple virtual depth layers based on their three-dimensional height coordinates, and particle counting is performed independently for each layer.

[0013] The output includes a quantitative report containing the total number of particles, the number of discolored particles, and the rejection efficiency.

[0014] As an optional solution of the quantitative method for polystyrene microplastic particles described in this invention, the method involves: calculating the Euclidean color difference ΔE between the current dominant particle color and each color in the preset background color library, selecting the background color with the largest ΔE value, and completing the background color switching before the entire particle enters the imaging field of view.

[0015] Among them, the calculation of the European color difference ΔE is as follows:

[0016]

[0017] in, The color difference between two colors in the CIELAB color space;

[0018] , and The average color value of the grain region, and its coordinates in three dimensions of CIELAB space;

[0019] , and The coordinates in CIELAB space corresponding to the color value of the preset background color.

[0020] As an optional embodiment of the method for quantitatively measuring polystyrene microplastic particles according to the present invention, wherein the step of obtaining three-dimensional height information employs laser triangulation, including:

[0021] A laser line is projected onto the heterochromatic particles. The offset of the laser line on the particle surface is captured by a high-speed camera, and the Z-axis coordinate of the particles is calculated based on geometric relationships.

[0022]

[0023] in, The height of the particle's center point relative to the reference plane;

[0024] The number of pixels offset by the laser line is obtained through image processing;

[0025] The pixel size is the actual physical length represented by each pixel.

[0026] These are the fixed angles between the laser and the camera, and between the camera's optical axis and the vertical direction, respectively.

[0027] As an alternative to the method for quantitatively analyzing polystyrene microplastic particles according to the present invention, the method for identifying discolored particles includes:

[0028] A finely tuned YOLOv8-seg semantic segmentation model is used to generate granular pixel-level masks. Each mask region is clustered in the HSV color space. If the cluster center deviates from the preset target color threshold range, it is determined to be a heterochromatic particle.

[0029] As an alternative to the method for quantitatively measuring polystyrene microplastic particles according to the present invention, the method for predicting the landing point position includes:

[0030] State estimation of heterochromatic particles at positions above a certain number of consecutive frames is performed using a Kalman filter to obtain their velocity and acceleration components. These components are then substituted into the parabolic equation and horizontal displacement equation under gravity. The parabolic equation under gravity includes:

[0031]

[0032] in, The vertical height of the particle at time t is the Z-axis coordinate.

[0033] The initial height at the initial moment. Let be the initial vertical velocity at the initial moment. For time variables, The distance of fall caused by gravity;

[0034] The horizontal displacement equations include:

[0035]

[0036] in, The horizontal position of the particle at time t is represented by its X-axis coordinate.

[0037] The initial horizontal starting position at the initial moment. The initial horizontal velocity at the initial moment;

[0038] The landing coordinates after leaving the vibration platform are calculated based on the parabolic equation and the horizontal displacement equation.

[0039] As an optional embodiment of the quantitative method for polystyrene microplastic particles described in this invention, when the predicted landing point is located in the center qualified area, an instantaneous acceleration pulse in the Z-axis direction is applied just before the particle leaves the platform, and the platform tilt angle is adjusted synchronously to change the direction of the initial velocity of the particle.

[0040] As an optional solution to the method for quantifying polystyrene microplastic particles described in this invention, the following steps are taken: at least two or more images with different focal planes are acquired, and after synthesizing a fully focused image, each particle is assigned to two equally spaced depth layers according to its Z-axis coordinate. Independent connected component analysis and area filtering are performed on each layer, and Laplacian gradient fusion and depth layer counting are used for comprehensive calculation to avoid counting errors caused by particle stacking.

[0041]

[0042] in, For each image at position ( The Laplacian gradient magnitude of the image is used to measure sharpness.

[0043] To obtain the pixel value at position (x, y) on the final synthesized full-focus image;

[0044] The image index with the largest gradient magnitude determines which pixel on the focal plane is selected for the final composite image.

[0045] As an optional solution of the method for quantifying polystyrene microplastic particles according to the present invention, a verification imaging unit is set downstream of the rejection area to confirm the rejection result in real time.

[0046] If discolored particles are detected but not successfully removed, a secondary removal instruction is triggered or a missed detection event is recorded, and the removal success rate is included in the quantitative report.

[0047] The present invention has the following beneficial effects:

[0048] 1. This method for quantitatively analyzing polystyrene microplastic particles automatically identifies the main color of the current batch of feed, eliminating the need for manual presets and significantly improving versatility. It also pre-determines the color, allowing sufficient time for subsequent background switching and effectively avoiding misjudgments caused by uneven lighting or surface reflection. Based on the target color, it automatically switches the imaging background to the color with the highest contrast to enhance the visual distinguishability of dissimilar particles.

[0049] 2. This method for quantifying polystyrene microplastic particles significantly improves image quality by enhancing the edge contrast between particles and background through dynamic background and ensures consistent imaging under all weather conditions through ambient light adaptive capability.

[0050] 3. This method for quantifying polystyrene microplastic particles achieves micron-level three-dimensional positioning in high-speed online scenarios, providing crucial depth information for subsequent trajectory prediction and layered counting. It fundamentally solves the problem that traditional two-dimensional imaging cannot distinguish overlapping particles in upper and lower layers, laying the foundation for high-precision measurement. Attached Figure Description

[0051] Figure 1 This is a schematic diagram of the pre-identification process of the present invention.

[0052] Figure 2 This is a schematic diagram illustrating the trajectory prediction and targeted elimination process of the present invention.

[0053] Figure 3 This is a schematic diagram of the hierarchical calculation and result verification process of the present invention. Detailed Implementation

[0054] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] Example 1

[0056] Please see Figures 1-3 One method for quantifying polystyrene microplastic particles includes:

[0057] Real-time acquisition of images of particles during transmission, and identification of the dominant target color;

[0058] Specifically, approximately 20 to 30 milliseconds before particles enter the main imaging area, the system activates a low-resolution pre-scan camera, such as 640×480 pixels, at a frame rate of no less than 200 frames per second, to capture the initial image. Through connected component analysis, all particle regions are extracted from the image and converted from the RGB color space to the HSV color space. The hue and saturation of each region are then statistically analyzed. If a certain color category, such as the standard white of undyed polystyrene, has a hue between 0 and 10 degrees or 350 and 360 degrees, a saturation below 0.15, and occupies more than 80% of the field of view, then this color is determined to be the dominant target color. To enhance robustness, a lightweight neural network, such as MobileNetV2, can be used for verification. By automatically identifying the dominant color of the current batch of feed material, no manual preset is required, significantly improving versatility. Simultaneously, color discrimination is completed in advance, reserving sufficient time for subsequent background switching and effectively avoiding misjudgments caused by uneven lighting or surface reflection.

[0059] The imaging background is automatically switched to the color with the highest contrast based on the target color in order to enhance the visual distinguishability of heterochromatic particles.

[0060] Specifically, a standard background color library based on the CIELAB color space is built-in, including black, white, red, green, and blue. After converting the identified dominant target color into CIELAB coordinates, the Euclidean color difference ΔE between it and each background color in the color library is calculated, and the color with the largest ΔE is selected as the current optimal background color. Control commands are sent to the RGBW LED backlight panel via a high-speed digital interface, such as SPI, and color switching is completed within 1 millisecond. At the same time, a coaxial ring LED supplementary light source is activated, with an illumination angle of 25 to 35 degrees, to suppress specular reflection on the surface of polystyrene particles. Dynamic background enhancement improves the edge contrast between particles and background, significantly improving image quality, and ambient light adaptive capability ensures consistent imaging in all weather conditions.

[0061] This involves identifying the dominant target particle color in the current batch through pre-scanned images. Specifically, connected component analysis is performed on the low-resolution image before it enters the field of view to extract all particle regions, and their average color value in the CIELAB color space is calculated, denoted as . , and , where represents the lightness, red-green axis (a*), and yellow-blue axis (b*) coordinates of the dominant particle color in CIELAB space, respectively. Then, this dominant color is compared with each background color in a preset background color library. The background color library contains several standard colors (such as black, white, red, green, and blue), each with a fixed coordinate value in CIELAB space, denoted as . , and For each background color, the system calculates the Euclidean color difference ΔE between it and the dominant particle color:

[0062]

[0063] in, The color difference between two colors in the CIELAB color space;

[0064] , and The average color value of the grain region, and its coordinates in three dimensions of CIELAB space;

[0065] , and The coordinates of the preset background color value in CIELAB space;

[0066] in, This represents the perceived color difference between two colors in the CIELAB color space. The larger the value, the more obvious the color difference perceived by the human eye, and the higher the visual contrast. By traversing all candidate background colors in the background color library, the corresponding ΔE value is calculated, and the background color with the largest ΔE is selected as the current optimal imaging background. Finally, before the overall grain officially enters the high-resolution main imaging field of view, a switching command is sent to the RGBWLED backlight panel through a high-speed control interface, such as SPI or DMX512, to complete the real-time switching of the background color.

[0067] Color difference calculations are performed based on the CIELAB color space, which is perceived as uniform by the human eye, to ensure that the selected background color has the greatest visual contrast with the dominant grain color; by switching before the grain enters the main field of view, image blurring caused by background flickering or delay during the imaging process is avoided.

[0068] Simultaneously acquire the two-dimensional position information and three-dimensional height information of the particles to construct the spatial state vector of each particle;

[0069] Specifically, after the background transition, the high-resolution main camera and the 650-nanometer-wire laser work together. The main camera captures color images of the particles to obtain their XY-plane positions, while the high-speed auxiliary camera simultaneously records the offset deformation of the laser line on the particle surface. Using the principle of laser triangulation, combined with calibrated geometric parameters, a laser incident angle of 45 degrees, a camera viewing angle of 30 degrees, and a pixel size of 3.45 micrometers / pixel, the Z-coordinate of the top of the particle is calculated with an accuracy of ±2 micrometers. Finally, the spatial state vector of each particle is defined as a six-dimensional vector [x, y, z, v_x, v_y, v_z], where the velocity component is initially estimated by multi-frame position difference. By achieving micrometer-level three-dimensional positioning in a high-speed online scene, key depth information is provided for subsequent trajectory prediction and layer counting, fundamentally solving the problem that traditional two-dimensional imaging cannot distinguish between overlapping particles in upper and lower layers, laying the foundation for high-precision measurement.

[0070] Based on color classification, non-target color particles are identified and assigned a unique identifier.

[0071] Specifically, a finely tuned YOLOv8-seg semantic segmentation model is used to perform pixel-level instance segmentation on high-resolution images. For each segmentation mask, the HSV values ​​of all pixels are extracted, and the dominant color center is obtained through K-means clustering. If the saturation is greater than 0.2 or the hue is not within the white range, it is identified as a heterochromatic particle. The system assigns it a globally unique ID, such as R023 representing the 23rd red particle, and establishes a mapping table between IDs and states in memory, supporting cross-frame tracking. This pixel-level segmentation avoids the omission or over-segmentation of small particles by traditional bounding box methods. HSV clustering is not sensitive to changes in illumination and maintains high classification accuracy under complex lighting conditions. The unique ID mechanism supports the simultaneous tracking of multiple heterochromatic particles, preventing misplacement during elimination.

[0072] The motion trajectory of heterochromatic particles is tracked using multiple consecutive frames of images, and their parabolic motion model is fitted to predict the impact point.

[0073] A narrow laser line with a wavelength of 650 nanometers is projected onto a heterochromatic particle during transmission. The laser line is reflected off the particle surface, and its offset is proportional to the height of the particle.

[0074] A laser line is projected onto the heterochromatic particles. The offset of the laser line on the particle surface is captured by a high-speed camera, and the Z-axis coordinate of the particles is calculated based on geometric relationships.

[0075]

[0076] in, The height of the particle's center point relative to the reference plane;

[0077] The number of pixels offset by the laser line is obtained through image processing;

[0078] The pixel size is the actual physical length represented by each pixel.

[0079] These are the fixed angles between the laser and the camera, and between the camera's optical axis and the vertical direction, respectively.

[0080] Employing a high-resolution camera and sub-pixel edge detection technology, combined with a precise geometric model, it can achieve micron-level height measurement accuracy, which is crucial for analyzing tiny particles. The 3D height information not only improves the accuracy of segmentation and recognition but also provides necessary data support for advanced functions such as trajectory prediction and virtual layer counting. It solves the problem that traditional 2D imaging cannot distinguish between overlapping particles in upper and lower layers. By optimizing the laser incident angle and camera viewpoint, the influence of ambient light and other optical noise can be effectively reduced, ensuring the consistency and reliability of the measurement results.

[0081] Specifically, for each heterochromatic particle ID, an independent Kalman filter is maintained, fusing position observations from multiple consecutive frames to output a smooth trajectory. When a particle approaches the platform edge, the vertical motion equation is solved:

[0082]

[0083] in, The vertical height of the particle at time t is the Z-axis coordinate.

[0084] The initial height at the initial moment. Let be the initial vertical velocity at the initial moment. For time variables, The distance of fall caused by gravity;

[0085] Let z(t) = 0 to solve for the landing time t_f, then substitute it into the horizontal displacement equation to obtain the predicted landing point coordinates;

[0086] The horizontal displacement equations include:

[0087]

[0088] in, The horizontal position of the particle at time t is represented by its X-axis coordinate.

[0089] The initial horizontal starting position at the initial moment. The initial horizontal velocity at the initial moment;

[0090] Calculate its landing coordinates after leaving the vibration platform based on the parabolic equation and the horizontal displacement equation.

[0091] The vibration parameters of the vibration platform and the direction of the edge airflow are dynamically adjusted according to the predicted landing point to direct the discolored particles to the rejection area.

[0092] Specifically, if the predicted landing point is located in the center qualified zone, the system triggers a combined action before the particle leaves the stage: the Z-axis piezoelectric actuator applies a 6g instantaneous acceleration pulse, the servo motor tilts the platform 1.2 degrees to the right, and simultaneously the right-side micro-nozzle outputs a 6 m / s directional airflow. These three actions work together to shift the particle's landing point at least 15 mm to the right, ensuring it falls into the rejection slot.

[0093] The particles are assigned to multiple virtual depth layers based on their three-dimensional height coordinates, and particle counting is performed independently for each layer.

[0094] Specifically, synchronous focus stacking imaging is performed, with the Z-axis platform driving the camera to capture images at five focal planes (0, 10, 20, 30, and 40 micrometers). A full-focus image is generated through Laplacian gradient fusion. Based on the Z-coordinate, particles are assigned to five equally thick virtual layers (8 micrometers each). Connectivity analysis is performed independently on each layer, and filtering is performed based on area and circularity (greater than 0.7), outputting the layer counting results.

[0095] A Z-axis precision displacement platform or motorized lens is controlled to acquire images of at least two (preferably five) different focal planes in the particle imaging area. These focal planes are evenly spaced along the Z-axis, for example, located at heights of 0 μm, 10 μm, 20 μm, 30 μm, and 40 μm, respectively, covering the thickness range of a typical microplastic particle.

[0096] Subsequently, for each focal plane image Ii(x, y) (where i = 1, 2, ..., N, N ≥ 2), the Laplacian gradient magnitude Gi(x, y) at each pixel location (x, y) is calculated to measure the local sharpness at that location. The formula for calculating the Laplacian gradient magnitude is:

[0097]

[0098] in, This represents the second-order Laplace operator; the larger the value, the more focused the region is.

[0099] Next, the system synthesizes a single panfocus image. (x, y), where each pixel value is taken from the image with the largest gradient magnitude at that location among all focal plane images:

[0100]

[0101] in, For each image at position ( The Laplacian gradient magnitude of the image is used to measure sharpness.

[0102] To obtain the pixel value at position (x, y) on the final synthesized full-focus image;

[0103] That is, k is the image index with the largest Laplacian gradient magnitude at position (x, y), which determines which focal plane the pixel should be selected from in the final synthesized image.

[0104] While acquiring a full-focus image, the system obtains the precise Z-axis coordinates of the center point of each particle using laser triangulation. Based on these Z-coordinates, all particles are assigned to two or more pre-defined, equally spaced depth layers (e.g., Layer 1: [0, 20] μm, Layer 2: [20, 40] μm). Connectivity analysis is performed independently on each layer, combined with area filtering (e.g., retaining only connected regions of 50-10,000 μm²) and shape constraints (e.g., roundness > 0.7) to remove noise or debris interference. Finally, the system counts the number of particles in each depth layer and sums them to obtain the total number of particles, thus achieving high-precision quantification.

[0105] The output includes a quantitative report containing the total number of particles, the number of discolored particles, and the rejection efficiency.

[0106] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0107] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for quantitatively measuring polystyrene microplastic particles, characterized in that, include: Real-time acquisition of images of particles during transmission, and identification of the dominant target color; The imaging background is automatically switched to the color with the highest contrast based on the target color to enhance the visual distinguishability of dissimilar particles. Simultaneously acquire the two-dimensional position information and three-dimensional height information of the particles to construct the spatial state vector of each particle; Based on color classification, non-target color particles are identified and assigned a unique identifier. The motion trajectory of the heterochromatic particles is tracked using multiple consecutive images, and their parabolic motion model is fitted to predict the impact point. The vibration parameters of the vibration platform and the direction of the edge airflow are dynamically adjusted according to the predicted landing point to directionally throw the discolored particles to the rejection area. The particles are assigned to multiple virtual depth layers based on their three-dimensional height coordinates, and particle counting is performed independently for each layer. The output includes a quantitative report containing the total number of particles, the number of discolored particles, and the rejection efficiency.

2. The method for quantitatively measuring polystyrene microplastic particles according to claim 1, characterized in that: Calculate the Euclidean color difference ΔE between the current dominant particle color and each color in the preset background color library, and select the background color with the largest ΔE value to complete the background color switching before the entire particle enters the imaging field of view. Among them, the calculation of the European color difference ΔE is as follows: in, The color difference between two colors in the CIELAB color space; , and The average color value of the grain region, and its coordinates in three dimensions of CIELAB space; , and The coordinates in CIELAB space corresponding to the color value of the preset background color.

3. The method for quantitatively measuring polystyrene microplastic particles according to claim 2, characterized in that, The step of obtaining three-dimensional height information uses laser triangulation, including: A laser line is projected onto the heterochromatic particles. The offset of the laser line on the particle surface is captured by a high-speed camera, and the Z-axis coordinate of the particles is calculated based on geometric relationships. in, The height of the particle's center point relative to the reference plane; The number of pixels offset by the laser line is obtained through image processing; The pixel size is the actual physical length represented by each pixel. These are the fixed angles between the laser and the camera, and between the camera's optical axis and the vertical direction, respectively.

4. The method for quantitatively measuring polystyrene microplastic particles according to claim 3, characterized in that, The method for identifying discolored particles includes: A finely tuned YOLOv8-seg semantic segmentation model is used to generate granular pixel-level masks. Each mask region is clustered in the HSV color space. If the cluster center deviates from the preset target color threshold range, it is determined to be a heterochromatic particle.

5. The method for quantitatively measuring polystyrene microplastic particles according to claim 4, characterized in that, The method for predicting the landing point location includes: State estimation of heterochromatic particles at positions above a certain number of consecutive frames is performed using a Kalman filter to obtain their velocity and acceleration components. These components are then substituted into the parabolic equation and horizontal displacement equation under gravity. The parabolic equation under gravity includes: in, The vertical height of the particle at time t is the Z-axis coordinate. The initial height at the initial moment. Let be the initial vertical velocity at the initial moment. For time variables, The distance of fall caused by gravity; The horizontal displacement equations include: in, The horizontal position of the particle at time t is represented by its X-axis coordinate. The initial horizontal starting position at the initial moment. The initial horizontal velocity at the initial moment; The landing coordinates after leaving the vibration platform are calculated based on the parabolic equation and the horizontal displacement equation.

6. The method for quantitatively measuring polystyrene microplastic particles according to claim 5, characterized in that: When the predicted landing point is in the center qualified zone, an instantaneous acceleration pulse in the Z-axis direction is applied just before the particle leaves the platform, and the platform tilt angle is adjusted simultaneously to change the direction of the initial velocity of the particle.

7. The method for quantitatively measuring polystyrene microplastic particles according to claim 6, characterized in that: At least two images with different focal planes are acquired, and after synthesizing the full-focus image, each particle is assigned to two equally spaced depth layers according to its Z-axis coordinate. Independent connected component analysis and area filtering are performed on each layer. Laplacian gradient fusion and depth layer counting are used for comprehensive calculation to avoid counting errors caused by particle stacking. in, For each image at position ( The Laplacian gradient magnitude of the image is used to measure sharpness. To obtain the pixel value at position (x, y) on the final synthesized full-focus image; The image index with the largest gradient magnitude determines which pixel on the focal plane is selected for the final composite image.

8. The method for quantitatively measuring polystyrene microplastic particles according to claim 7, characterized in that: A verification imaging unit is set up downstream of the removal area to confirm the removal results in real time; If discolored particles are detected but not successfully removed, a secondary removal instruction is triggered or a missed detection event is recorded, and the removal success rate is included in the quantitative report.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-8.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-8.

Images