A method and system for detecting PE microplastics from soil mulch film
By combining visible light and fluorescence images for initial screening and segmentation, the detection of microplastics in soil is achieved, solving the problems of slow detection speed, high cost and high false positive rate in existing technologies, thus realizing efficient and accurate microplastic detection.
Patent Information
- Application Number
- CN202611399259.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-10
- Publication Date
- 2026-11-17
AI Technical Summary
Existing microplastic detection methods in soil suffer from problems such as slow detection speed, high cost, high false positive rate, and high system complexity. In particular, it is difficult to accurately establish spatial correspondence between different imaging devices or detection stations, which affects the accuracy of detection.
By acquiring visible light and fluorescence images of the soil surface, preliminary screening of suspected plastic film areas and segmentation of fluorescence areas are performed. Combined with spatial intersection operation and identification of plastic film source PE, the detection of microplastics is achieved.
It significantly improves the accuracy of microplastic detection, reduces false positive and false negative rates, adapts to different soil types and light conditions, and provides a fast and convenient quantitative detection solution.
Smart Images

Figure CN123111965A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of microplastic detection technology, specifically to a method and system for detecting PE microplastics from soil mulch film. Background Technology
[0002] Mulching technology is widely used in modern agricultural production, significantly improving crop yields and planting efficiency. However, polyethylene (PE) mulch film remaining in the soil gradually breaks down into microplastic particles smaller than 5 mm in diameter under long-term natural weathering, mechanical tillage, and ultraviolet radiation. These PE microplastics from mulch film accumulate in the soil year by year, not only altering the soil's physical structure but also potentially posing a threat to the ecological environment and human health through crop absorption or transmission via the food chain. Therefore, achieving rapid and accurate detection of PE microplastics from mulch film in the soil has become a crucial problem urgently needing to be solved in the field of agricultural environmental monitoring.
[0003] Existing microplastic detection methods are mainly divided into two categories: spectroscopic analysis and morphological visual methods. While spectroscopic methods (such as Fourier transform infrared spectroscopy (FTIR) and Raman spectroscopy) can provide qualitative and quantitative information on the composition of substances, the equipment is expensive, sample preparation is cumbersome, and single-point measurements are time-consuming, making it difficult to meet the needs of rapid screening of large batches of soil samples. Morphological visual methods identify microplastics based on the appearance features such as color and shape in visible light images. They are fast and relatively inexpensive, but the numerous mineral particles, organic matter aggregates, and crop residues present in the soil matrix are highly similar in color and shape to plastic film fragments, making misjudgment based solely on morphological features highly likely, resulting in a high false positive rate.
[0004] Meanwhile, PE microplastics derived from plastic film can produce characteristic fluorescence emission under ultraviolet light excitation (especially in the 350-380 nm band), while the fluorescence signals of background interference such as soil minerals and humus are weaker or have different spectral characteristics. This difference provides a potential basis for the targeted identification of PE based on chemical specificity. However, ultraviolet fluorescence images lack the morphological details of visible light images, making it difficult to distinguish microplastics from fluorescent impurities in the soil; while visible light images are rich in morphological information, they cannot provide a basis for distinguishing chemical composition. How to effectively integrate the information advantages of the two modalities of images, while overcoming the problems of insufficient accuracy and high computational cost of traditional image registration algorithms when processing loose soil surfaces, is the core challenge currently facing the technology.
[0005] In addition, existing joint spectral and morphological detection schemes usually require transferring samples between different imaging devices or different detection stations, which not only increases the system complexity and detection time, but also makes it difficult to accurately establish the spatial correspondence between visible light images and fluorescence images due to sample position changes, affecting the accuracy of subsequent fusion and discrimination. Summary of the Invention
[0006] This application aims to provide a method and system for detecting PE microplastics from soil mulch film sources, which can improve the accuracy of microplastic detection.
[0007] The technical solution of this application is implemented as follows: In a first aspect, embodiments of this application provide a method for detecting PE microplastics from soil mulch film sources, the method comprising: Acquire visible light images, raw fluorescence images, and background fluorescence images of the soil surface; Based on the visible light image of the soil surface, a preliminary screening of suspected plastic film areas is performed to obtain suspected visible light masks. Based on the original fluorescence image and the background fluorescence image, fluorescence region segmentation is performed to obtain a fluorescence region mask; Based on the visible light suspected mask and the fluorescent region mask, spatial intersection operation and PE source discrimination of the mulch film are performed to obtain microplastic detection results; and the microplastic detection results are converted into soil abundance.
[0008] In the above scheme, the step of performing preliminary screening of suspected plastic film areas based on the visible light image of the soil surface to obtain a suspected visible light mask includes: The visible light image of the soil surface is preprocessed to obtain a hue channel image, a saturation channel image, and a brightness channel image; Based on the hue channel image and the saturation channel image, a coarse screening of color and brightness thresholds is performed to obtain a primary candidate mask; Based on the brightness channel image, the local standard deviation of brightness values in the neighborhood of each pixel is calculated to generate a texture contrast map; and based on the texture contrast map and a preset contrast upper limit threshold, a texture filtering mask is determined. The primary candidate mask and the texture screening mask are fused and morphologically modified to obtain the visible light potential mask.
[0009] In the above scheme, the preprocessing of the visible light image of the soil surface to obtain the hue channel image, saturation channel image, and brightness channel image includes: The visible light image of the soil surface is smoothed by a smoothing filter to obtain a smoothed visible light image; Spatial mapping is performed on the smoothed visible light image to obtain the hue channel image, the saturation channel image, and the brightness channel image.
[0010] In the above scheme, the step of segmenting the fluorescence region based on the original fluorescence image and the background fluorescence image to obtain a fluorescence region mask includes: The original fluorescence image is subjected to pixel-level background correction using the background fluorescence image to obtain the corrected fluorescence image; Based on the corrected fluorescence image, fluorescence region segmentation is performed to obtain the fluorescence region mask.
[0011] In the above scheme, the step of segmenting the fluorescence region based on the corrected fluorescence image to obtain the fluorescence region mask includes: The corrected fluorescence image is subjected to Gaussian smoothing and noise reduction processing to obtain a smoothed fluorescence image; Based on the smoothed fluorescence image, a global grayscale histogram is calculated; and a preliminary screening fluorescence mask is generated according to the global grayscale histogram and a preset optimal segmentation threshold. By setting a minimum effective fluorescence area threshold, the initial screening fluorescence mask is denoised by area thresholding to obtain the fluorescence region mask.
[0012] In the above scheme, the step of performing spatial intersection calculation and PE source discrimination based on the visible light suspected mask and the fluorescent region mask to obtain microplastic detection results includes: Perform spatial intersection operation and effective target marking on the visible light suspected mask and the fluorescent region mask to obtain the spatial intersection mask and the total number of effective targets corresponding to the spatial intersection mask; For each valid target in the spatial intersection mask, joint feature extraction is performed based on the spatial intersection mask, the original fluorescence image, and the visible light image to obtain the joint features of each valid target; Calculate the comprehensive PE suspicion score for each valid target based on the joint features of each valid target; By combining the comprehensive PE suspected score of each valid target with a preset dynamic threshold, the PE source of the mulch film is identified, and the microplastic detection result is determined.
[0013] In the above scheme, the step of performing joint feature extraction on each effective target in the spatial intersection mask based on the spatial intersection mask, the original fluorescence image, and the visible light image to obtain the joint features of each effective target includes: For each valid target in the spatial intersection mask, fluorescence features are extracted based on the original fluorescence image and the spatial intersection mask to obtain a fluorescence feature group for each valid target; Based on the spatial intersection mask and the visible light image, morphological features are extracted to obtain a morphological feature group for each effective target; A joint feature of each effective target is generated by combining the fluorescence feature set and the morphological feature set of each effective target.
[0014] Secondly, embodiments of this application provide a soil mulch film-derived PE microplastic detection system. This system includes: an acquisition module, a screening module, a segmentation module, and a detection module; wherein... The acquisition module is used to acquire visible light images of the soil surface, raw fluorescence images, and background fluorescence images; The screening module is used to perform preliminary screening of suspected plastic film areas based on the visible light image of the soil surface to obtain suspected visible light masks; The segmentation module is used to segment the fluorescence region based on the original fluorescence image and the background fluorescence image to obtain a fluorescence region mask; The detection module is used to perform spatial intersection calculation and PE source discrimination based on the visible light suspected mask and the fluorescent region mask to obtain microplastic detection results; and convert the microplastic detection results into soil abundance.
[0015] Thirdly, embodiments of this application provide a soil mulch film-derived PE microplastic detection device, comprising: a processor and a memory; wherein, Memory, used to store computer programs; A processor is used to retrieve and run computer programs from memory to perform methods such as those described in the first aspect.
[0016] Fourthly, embodiments of this application provide a computer-readable storage medium storing executable instructions for causing a processor to perform the method as described in the first aspect.
[0017] This application provides a method and system for detecting PE microplastics from soil mulch film. The method includes: acquiring a visible light image, a raw fluorescence image, and a background fluorescence image of the soil surface; performing preliminary screening of suspected mulch film areas based on the visible light image of the soil surface to obtain a visible light suspected mask; performing fluorescence region segmentation based on the raw fluorescence image and the background fluorescence image to obtain a fluorescence region mask; performing spatial intersection operation and mulch film-source PE discrimination based on the visible light suspected mask and the fluorescence region mask to obtain microplastic detection results; and converting the microplastic detection results into soil abundance. In the above scheme, by performing preliminary screening of suspected film areas in the acquired visible light image of the soil surface, the effective range of subsequent fluorescence analysis can be significantly narrowed, effectively eliminating the interference of a large number of non-film mineral particles and organic matter aggregates in the soil, reducing the burden of subsequent fluorescence segmentation and discrimination; by performing fluorescence region segmentation on the acquired raw fluorescence image and background fluorescence image, automatic extraction of fluorescence regions without manually setting a fixed threshold is achieved, which has good adaptability to different soil types and lighting conditions and improves the detection sensitivity of weak fluorescence targets. By performing spatial intersection calculations and identifying PE sources from mulch film between visible light suspected masks and fluorescent area masks, the complementary advantages of the chemical specific information provided by the fluorescence signal and the physical morphological information of mulch film fragments provided by the morphological features can be fully utilized. This overcomes the limitations of single-mode or single-feature discrimination, significantly reduces the false detection rate and false negative rate, and thus improves the accuracy of microplastic detection. Attached Figure Description
[0018] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application. Obviously, the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.
[0019] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.
[0020] Figure 1 This is a schematic diagram of an optional process for a method of detecting PE microplastics from soil mulch film provided in an embodiment of this application; Figure 2 A schematic diagram of the initial screening binary image construction for a soil mulch film source PE microplastic detection method provided in this application embodiment; Figure 3A schematic diagram of a soil mulch film source PE microplastic detection system provided in this application embodiment; Figure 4 This is a schematic diagram of the structure of a soil mulch film source PE microplastic detection device provided in an embodiment of this application. Detailed Implementation
[0021] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the specific technical solutions of this application will be further described in detail below with reference to the accompanying drawings of the embodiments of this application. The following embodiments are used to illustrate this application, but are not intended to limit the scope of this application.
[0022] Unless otherwise defined, all technical and scientific terms used in this application have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this application is for the purpose of describing embodiments of this application only and is not intended to be limiting of this application.
[0023] In the following description, references to "some embodiments," "this embodiment," "this application embodiment," and examples, etc., describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same subset or different subset of all possible embodiments and may be combined with each other without conflict.
[0024] If the application documents contain similar descriptions such as "first / second", the following explanation shall be added: In the following description, the terms "first / second / third" are used only to distinguish similar objects and do not represent a specific order of objects. It is understood that "first / second / third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.
[0025] This application provides a method for detecting PE microplastics from soil mulch film. Figure 1 This is a schematic diagram of an optional process for detecting PE microplastics from soil mulch film, provided in an embodiment of this application. Figure 1 The steps shown are explained.
[0026] S101. Acquire visible light images, raw fluorescence images, and background fluorescence images of the soil surface.
[0027] In some embodiments of this application, under standardized lighting conditions (i.e., constant global illuminance, using a ring LED light source or sunlight), an industrial area array camera (i.e., a visible light camera) is used to acquire 24-bit true-color RGB images of the soil surface, i.e., visible light images of the soil surface, denoted as img_rgb (size is W_img×H_img).
[0028] In some embodiments of this application, the visible light camera and the ultraviolet fluorescence imaging system share the same optical axis / the same image sensor (or achieve a common field of view through a fixed beam splitting path), and the soil sample remains stationary after the visible light image is acquired. The spatial correspondence between the two imaging fields of view is fixed by the hardware structure, and no additional image registration is required at the algorithm level.
[0029] In some embodiments of this application, the visible light source is turned off (i.e., the ring LED light source is turned off or sunlight is blocked), and the ultraviolet light source (excitation band of 350~380 nm, preferably 365 nm) is turned on to irradiate the sample. A long-pass filter component is inserted in front of the camera lens (the cutoff wavelength is determined according to the target fluorescence emission band; in this embodiment, it can be 400~420 nm) to allow the target fluorescence signal to pass through and block the scattering components of the ultraviolet excitation light. The raw fluorescence image under ultraviolet excitation is acquired and recorded as img_uv_raw (resolution of W_img×H_img, grayscale image, bit depth of 16-bit, to preserve the dynamic range of weak signals).
[0030] In some embodiments of this application, the ultraviolet light source intensity, camera gain, and exposure time are kept exactly the same as described above. Under the condition of not placing soil samples (or placing a blank control of pure quartz sand confirmed to be non-fluorescent by high-temperature burning), background fluorescence images are acquired and denoted as img_bg (resolution is W_img×H_img); wherein, img_bg characterizes the dark current of the optical system, lens stray light, and fixed pattern noise in the environment.
[0031] S102. Based on the visible light image of the soil surface, perform preliminary screening of suspected plastic film areas to obtain suspected visible light masks.
[0032] In some embodiments of this application, S102 can be implemented by S1021-S1024, as follows: S1021. Preprocess the visible light image of the soil surface to obtain the hue channel image, saturation channel image, and brightness channel image.
[0033] In some embodiments of this application, a visible light image of the soil surface is smoothed by filtering to obtain a smoothed visible light image; the smoothed visible light image is then spatially mapped to obtain a hue channel image, a saturation channel image, and a brightness channel image.
[0034] For example, a median filter is applied to the visible light image `img_rgb` of the soil surface to remove salt-and-pepper noise, resulting in a smoothed visible light image `img_smooth`. `img_smooth` is then converted from the RGB color space to the HSV color space to obtain a hue channel image `h_channel`, a saturation channel image `s_channel`, and a lightness channel image `v_channel`, which are used for subsequent color and lightness feature extraction. The resolution of `h_channel`, `s_channel`, and `v_channel` is all `W_img×H_img`.
[0035] S1022. Based on the hue channel image and saturation channel image, perform coarse screening of color and brightness thresholds to obtain primary candidate masks.
[0036] S1023. Based on the brightness channel image, calculate the local standard deviation of brightness values in the neighborhood of each pixel to generate a texture contrast map; and determine the texture screening mask based on the texture contrast map and the preset contrast upper limit threshold.
[0037] S1024. Perform mask fusion and morphological trimming on the primary candidate mask and texture screening mask to obtain a visible light potential mask.
[0038] For example, the lower limit H_min and upper limit H_max of the common color range of mulch film are set (based on h_channel), the lower limit S_min and upper limit S_max of saturation are set (based on s_channel), and the lower limit V_min and upper limit V_max of brightness are set (based on v_channel, used to remove soil shadows and overexposed highlights). H_min is 20, H_max is 160, S_min is 40, S_max is 255, V_min is 25, and V_max is 230.
[0039] The process of generating the initial binary image is as follows: Figure 2 As shown, a logical AND operation is performed pixel by pixel to generate H_mask, S_mask and V_mask respectively.
[0040] If the pixel hue satisfies H_mask[i,j]∈[H_min,H_max], then assign H_mask[i,j]=1; otherwise, assign 0. S_mask and V_mask use the same judgment rule. Take the intersection of the three types of masks to generate a primary candidate mask mask_prim (mask_prim=H_mask∩S_mask∩V_mask) with the same resolution as the original image (W_img×H_img); where i and j are the corresponding image row and column coordinates.
[0041] The v_channel is used to replace the grayscale map, and local texture contrast is calculated to distinguish smooth thin films from rough soil aggregates. A sliding window size win_tex (e.g., 15×15 pixels) is set, the local standard deviation of the lightness value in the neighborhood of each pixel is calculated as the contrast measurement standard, and a texture contrast map tex_map (with a resolution of W_img×H_img) is generated. An upper contrast threshold C_max (the thin film surface is smooth with a very small local standard deviation) is set to generate a texture filtering mask mask_tex. If tex_map[i,j]<C_max, mask_tex[i,j] is assigned a value of 1; otherwise, mask_tex[i,j] is assigned a value of 0. C_max may be 3.
[0042] The intersection of mask_prim and mask_tex is taken to obtain a primary screening binary map mask_final_raw=mask_prim∩mask_tex. To fill small holes inside the thin film caused by slight contamination or reflection, and connect adjacent fragments caused by tearing, a morphological closing operation (dilation followed by erosion) is performed on mask_final_raw, and a final suspected region mask mask_suspect (with a resolution of W_img×H_img) is output.
[0043] S103: Perform fluorescent region segmentation based on the original fluorescent image and the background fluorescent image to obtain a fluorescent region mask.
[0044] In some embodiments of the present application, S103 may be implemented through S1031 to S1032, as follows: S1031: Perform pixel-level background correction on the original fluorescent image using the background fluorescent image to obtain a corrected fluorescent image.
[0045] S1032: Perform fluorescent region segmentation based on the corrected fluorescent image to obtain a fluorescent region mask.
[0046] In some embodiments of the present application, Gaussian smoothing and noise reduction processing is performed on the corrected fluorescent image to obtain a smoothed fluorescent image; a global grayscale histogram is calculated based on the smoothed fluorescent image; and a primary screening fluorescent mask is generated according to the global grayscale histogram and a preset optimal segmentation threshold; area threshold denoising is performed on the primary screening fluorescent mask through a preset minimum effective fluorescent area threshold to obtain the fluorescent region mask.
[0047] Exemplarily, pixel-level background correction is performed on the original ultraviolet image through the background fluorescent image, while eliminating illumination non-uniformity (bright center and dark edges of the ultraviolet LED array) and fixed system noise. The corrected fluorescent image is defined as img_fluor (with a resolution of W_img×H_img, floating-point type), and the pixel-by-pixel calculation formula is as follows: ; in, The pixel value at coordinates (i,j) of the original ultraviolet image; Let be the pixel value of the background fluorescence image at coordinates (i,j); The pixel value at coordinates (i,j) is the corrected fluorescence image. Very small positive numbers (such as 10) -6 This is used to prevent division by zero errors caused by a zero denominator. The output value of this formula is dimensionless and represents the gain relative to the background fluorescence intensity; a larger value indicates a more significant fluorescence response in the target region.
[0048] Perform the following sub-steps on the corrected fluorescence image img_fluor to generate a binarized fluorescence mask: Gaussian smoothing noise reduction: Two-dimensional Gaussian filtering is performed on img_fluor (in this example, the kernel size can be 5×5 and the standard deviation σ=1.5) to suppress high-frequency noise from the sensor and output a smoothed fluorescence image img_fluor_gau (resolution W_img× H_img).
[0049] Adaptive threshold segmentation: A global grayscale histogram is calculated for the smoothed fluorescence image `img_fluor_gau`. The OTSU (Otsu method) is used to automatically determine the preset optimal segmentation threshold `thr_otsu`, which can be 130. A preliminary fluorescence mask is generated pixel-by-pixel. If `img_fluor_gau[i,j] > thr_otsu`, then `mask_fluor_raw[i,j] = 1`; otherwise, `mask_fluor_raw[i,j] = 0`.
[0050] Area threshold denoising: Pre-set the minimum effective fluorescence area threshold area_fluor_min_px (unit: pixel) 2 The value is determined by the system resolution and the lower limit of microplastic detection (e.g., area_fluor_min_px = 20 pixels). The system iterates through all connected components in mask_fluor_raw, calculating the pixel area of each component. Connected components with an area smaller than area_fluor_min_px are identified as isolated noise points, and the corresponding pixel value in mask_fluor_raw is set to 0. Connected components with an area greater than or equal to this threshold are retained. After the iteration is complete, the system outputs a fluorescent region mask mask_fluor (resolution W_img × H_img, binary image) after noise removal.
[0051] S104. Based on the visible light suspected mask and the fluorescent region mask, perform spatial intersection operation and PE source discrimination of the mulch film to obtain microplastic detection results; and convert the microplastic detection results into soil abundance.
[0052] In some embodiments of this application, effective targets with both thin film morphology and fluorescence response are extracted by pixel-by-pixel spatial intersection operation of visible light suspected mask and fluorescent region mask. Based on their joint feature vector, the suspected PE microplastic detection result is output by a discriminative model and finally converted into soil abundance.
[0053] Understandably, preliminary screening of suspected film regions in the acquired visible light images of the soil surface can significantly narrow down the effective range of subsequent fluorescence analysis, effectively eliminating interference from a large number of non-film mineral particles and organic matter aggregates in the soil, thus reducing the burden of subsequent fluorescence segmentation and discrimination. By segmenting the acquired original fluorescence images and background fluorescence images into fluorescence regions, automatic extraction of fluorescence regions without the need for manually setting fixed thresholds is achieved, demonstrating good adaptability to different soil types and lighting conditions, and improving the detection sensitivity of weakly fluorescent targets. Performing spatial intersection operations and identifying PE sources from the mulch film by combining suspected visible light masks with fluorescence region masks can fully utilize the complementary advantages of the chemical specificity information provided by the fluorescence signal and the fragment physical morphology information provided by the morphological features, overcoming the limitations of single-modality or single-feature discrimination, significantly reducing the false detection rate and false negative rate, thereby improving the accuracy of microplastic detection.
[0054] In some embodiments of this application, the spatial intersection operation and PE source discrimination based on the visible light suspected mask and the fluorescent region mask in S104 to obtain the microplastic detection result can be achieved through S1041-S1044, as follows: S1041. Perform spatial intersection operation and effective target marking on the visible light suspected mask and the fluorescent region mask to obtain the spatial intersection mask and the total number of effective targets corresponding to the spatial intersection mask.
[0055] S1042. For each effective target in the spatial intersection mask, perform joint feature extraction based on the spatial intersection mask, the original fluorescence image, and the visible light image to obtain the joint features of each effective target.
[0056] In some embodiments of this application, for each effective target in the spatial intersection mask, fluorescence features are extracted based on the original fluorescence image and the spatial intersection mask to obtain a fluorescence feature group for each effective target; morphological features are extracted based on the spatial intersection mask and the visible light image to obtain a morphological feature group for each effective target; and a joint feature for each effective target is generated through the fluorescence feature group and the morphological feature group of each effective target.
[0057] S1043. Calculate the comprehensive PE suspected score for each valid target based on the joint features of each valid target.
[0058] S1044. By using the comprehensive PE suspected score of each valid target and the preset dynamic threshold, the PE source of the mulch film is identified, and the microplastic detection results are determined.
[0059] For example, since the visible light image and the ultraviolet fluorescence image have already achieved pixel-level spatial alignment at the hardware level (both have a resolution of W_img×H_img and are in the same pixel coordinate system), a pixel-by-pixel logical AND (multiplication) operation is directly performed on the obtained visible light suspected mask mask_suspect and the obtained fluorescence region mask mask_fluor to generate a spatial intersection mask mask_overlap (with a resolution of W_img×H_img): mask_overlap[i,j]=mask_suspect[i,j]× mask_fluor[i,j]; Where mask_suspect[i,j] is the pixel value of the visible light suspected mask at coordinate (i,j); mask_fluor[i,j] is the pixel value of the fluorescent region mask at coordinate (i,j); mask_overlap[i,j] is the pixel value of the spatial intersection mask at coordinate (i,j); and pixels with a value of 1 in mask_overlap represent those locations that simultaneously satisfy both the visible light morphology suspected film and the ultraviolet fluorescence response. Eight-neighborhood connected component labeling is performed on mask_overlap, and the total number of valid targets M is counted. For the k-th connected component (k∈[0,M-1]), its pixel set is the joint region that simultaneously possesses both the visible light suspected film and fluorescence response; all subsequent features are extracted based on this intersection region.
[0060] For the kth valid target (k∈[0,M-1]), based on the kth connected component in the intersection mask_overlap, the original fluorescence intensity map img_fluor, and the visible light smoothing map img_smooth, the following two sets of features are extracted: Fluorescent signature (from img_fluor, applied to the k-th connected region of mask_overlap): fluor_mean_int[k]: The average grayscale value of all pixels in the k-th connected region on img_fluor.
[0061] fluor_area_px[k]: The pixel area of the k-th connected component (unit: pixel) 2 ), which is the pixel count of the kth connected component of mask_overlap.
[0062] fluor_integral[k]: Fluorescence intensity integral, calculated using the following formula: fluor_integral[k]=fluor_mean_int[k]×fluor_area_px[k]; Morphological Feature Group: The calculation of this group of features is uniformly based on the k-th connected component of the obtained mask_overlap and its corresponding pixel contour on the visible light smoothed image img_smooth. Before performing feature calculation, the following geometric quantities are first defined for use in this step: 1. perimeter[k]: The perimeter of the kth connected component (unit: pixel), calculated by a contour tracing algorithm (such as the Suzuki algorithm) from the binary contour of the connected component in img_smooth; 2. w_bbox[k], h_bbox[k]: The width and height (in pixels) of the smallest bounding rectangle of the k-th connected component, calculated by the rotating caliper algorithm on the set of contour points; 3. bbox_area[k]: The area of the smallest bounding rectangle of the k-th connected component (unit: pixels) 2 ), that is, w_bbox[k] ×h_bbox[k].
[0063] The above geometric quantities are all obtained by performing contour detection and minimum bounding rectangle calculation once on the contour point set of the k-th connected component of mask_overlap, without repeated traversal. Based on the above geometric quantities, the following four morphological features are extracted: asp_ratio[k]: Aspect ratio, calculated as the ratio of the major axis to the minor axis of the smallest circumscribed rectangle of the intersection region. Used to distinguish thin films (aspect ratio significantly > 1) from approximately isometric mineral grains. The calculation formula is: ; circularity[k]: This value is used to exclude round or near-round soil aggregates. A value closer to 1 indicates a shape closer to a circle. Film fragments are typically irregular strips or flakes, and this value is much less than 1. The calculation formula is: circularity[k]=4π*fluor_area_px[k] / (perimeter[k] 2 ); bnd_rough[k]: Boundary roughness is used to quantify the irregularity of the tear edge. The calculation formula is: bnd_rough[k] = perimeter[k] / (2 * (w_bbox[k] + h_bbox[k])) - 1, where this value is ≥ 0. The larger the value, the more severe the jagged tearing at the edge. Plastic film fragments often exhibit high roughness (typically > 0.05) due to mechanical breakage, while soil mineral particles have smooth edges, and this value is usually close to 0. deformation[k]: Deformation is defined as the ratio of the area of the smallest circumscribed rectangle of revolution to the actual area. The value is ≥1.0. The larger the value, the more severe the overall curling, bending or deformation of the fragments, which is consistent with the actual characteristics of aging and mechanical breakage of plastic film.
[0064] A comprehensive score is calculated for each extracted valid target, and its status as a PE microplastic source from mulch film is determined based on the score threshold.
[0065] First, the background mean fluorescence intensity bg_mean_int is calculated as a reference value for the soil background autofluorescence level. The average grayscale value of all pixels in img_fluor that do not belong to the mask_suspect region is then taken. bg_mean_int=mean{img_fluor[i,j]|mask_suspect[i,j]=0}; Subsequently, a comprehensive PE (Possible Detection and Presence) score is calculated for the k-th valid target (k∈[0,M-1]). The scoring function integrates five dimensions of characteristics: fluorescence response intensity, non-circularity, aspect ratio, boundary roughness, and deformation, with each dimension having a weight sum of 1. The comprehensive PE score is calculated as follows: ; Wherein, α is the fluorescence response intensity, the fluorescence gain of the target relative to the soil background; the higher the value, the more it tends to be PE (polymorphism), and the preferred value is 0.35; β is the non-circularity; the more the target deviates from a circle (i.e., the closer it is to the shape of a film fragment), the greater the contribution of this term, and the preferred value is 0.20; γ is the aspect ratio; the larger the aspect ratio of the target (the more elongated the shape), the greater the contribution of this term, but it is truncated by asp_ratio_max to prevent amplification of abnormal noise, and the preferred value is 0.15; δ is the boundary roughness; the more severe the edge jaggedness (reflecting mechanical fragmentation characteristics), the greater the contribution of this term, and the preferred value is 0.15; ζ is the deformation degree; the more severe the deviation of the target from a regular rectangle (reflecting curling / bending characteristics), the greater the contribution of this term, and the preferred value is 0.15; asp_ratio_max is a preset aspect ratio truncation value (e.g., asp_ratio_max = 20.0) to prevent a very small number of elongated noise points from causing abnormal amplification of this term.
[0066] thr_score is a dynamic threshold pre-determined by negative control soil experiments, which is usually set as the statistical mean of score_pe of negative control samples plus 3 times the standard deviation, to ensure a false positive rate < 0.3%.
[0067] The final determination logic is as follows: If score_pe[k]>=thr_score, the k-th target is determined as a mulch-derived PE microplastic, and label_pe[k] is marked as 1; If score_pe[k]<thr_score, the k-th target is determined as a non-PE interferent, label_pe[k] is marked as 0, and this target will not participate in subsequent abundance calculation.
[0068] Quantitative calculation is performed only for valid targets with label_pe[k]=1 (there are a total of count_pe=sum(label_pe) such targets).
[0069] First, mass conversion is performed. Standard PE samples with known masses (such as 0.1 mg, 0.5 mg, 1.0 mg, 2.0 mg and 5.0 mg) are used, their fluorescence images are collected under the same imaging conditions of the present method, the fluorescence integral intensity fluor_integral of each standard sample region is extracted, and a calibration relationship between PE mass and fluorescence integral intensity is established. In this embodiment, this calibration relationship can adopt a unary linear regression fitting passing through the origin, that is: mass_est_mg =k_calib×fluor_integral, where k_calib is the calibration coefficient (unit: mg / fluorescence integral intensity), obtained by linear regression fitting.
[0070] According to the above calibration coefficient, the equivalent mass mass_est_mg[k] (unit: mg) of the k-th target determined as suspected PE is calculated by the following formula: mass_est_mg[k]=k_calib*fluor_integral[k].
[0071] The masses of all targets determined as suspected PE are accumulated to obtain the total PE mass in the field of view: ; Meanwhile, the total soil mass soil_mass_g corresponding to the field of view of a single frame of image is calculated. The field of view area fov_cm2 (unit: cm 2 ) is pre-measured by a calibration plate, the sampling depth depth_cm is a fixed value (such as 0.5 cm), and the soil bulk density rho_soil (unit: g / cm 3 ) is pre-measured by the core cutter method, then: soil_mass_g=fov_cm2×depth_cm×rho_soil.
[0072] The final output includes the following two abundance indices: Mass abundance output (unit: mg / kg, i.e., milligrams of PE microplastics per kilogram of soil): conc_pe_mg_kg=mass_total_mg×1000 / soil_mass_g; Abundance output (unit: individuals / kg): conc_num_per_kg = count_pe × 1000 / soil_mass_g; where count_pe is the number of valid targets.
[0073] This application has the following beneficial effects: First, by utilizing the typical distribution pattern of mulch film fragments in the HSV color space and the surface smoothness texture characteristics characterized by the local standard deviation of the brightness channel, a preliminary screening of suspected film areas was completed in the visible light image. This significantly narrowed the effective range of subsequent fluorescence analysis, effectively eliminating the interference of a large number of non-film mineral particles and organic matter aggregates in the soil, and reducing the burden of subsequent fluorescence segmentation and discrimination.
[0074] Secondly, the interference of uneven UV illumination and fixed system noise on fluorescence signals was eliminated by the normalized background subtraction method. Combined with OTSU adaptive threshold segmentation and area threshold denoising, automatic and stable extraction of fluorescence regions was achieved without the need for manual setting of fixed thresholds. It has good adaptability to different soil types and lighting conditions and improves the detection sensitivity of weak fluorescence targets.
[0075] Furthermore, joint features of five dimensions—fluorescence response intensity, aspect ratio, circularity, boundary roughness, and deformation—are extracted simultaneously from the spatially aligned intersection region. PE determination is then performed using a weighted comprehensive scoring model. This fully leverages the complementary advantages of fluorescence signals providing chemically specific information and morphological features providing physical morphological information of mulch film fragments, overcoming the limitations of single-modality or single-feature discrimination and significantly reducing false positive and false negative rates.
[0076] Finally, by establishing a pre-defined relationship between PE mass and fluorescence integral intensity, target determination and abundance conversion can be completed simultaneously in the same detection process. It can output two core evaluation indicators, mass abundance (mg / kg) and number abundance (units / kg), without the need for additional weighing or counting steps. This provides a rapid, convenient, and information-complete quantitative detection solution for soil microplastic pollution monitoring, and has good prospects for promotion and application.
[0077] Based on the above embodiments of the method for detecting PE microplastics from soil mulch film, this application also provides a system for detecting PE microplastics from soil mulch film, such as... Figure 3 As shown, Figure 3 This is a schematic diagram of a soil mulch film-derived PE microplastic detection system provided in an embodiment of this application. The soil mulch film-derived PE microplastic detection system 3 includes: an acquisition module 301, a screening module 302, a segmentation module 303, and a detection module 304; wherein, The acquisition module 301 is used to acquire visible light images of the soil surface, original fluorescence images, and background fluorescence images; The screening module 302 is used to perform preliminary screening of suspected plastic film areas based on the visible light image of the soil surface to obtain suspected visible light masks; The segmentation module 303 is used to segment the fluorescence region based on the original fluorescence image and the background fluorescence image to obtain a fluorescence region mask; The detection module 304 is used to perform spatial intersection operation and PE source discrimination based on the visible light suspected mask and the fluorescent region mask to obtain microplastic detection results; and convert the microplastic detection results into soil abundance.
[0078] Based on the above embodiments of the method for detecting PE microplastics from soil mulch film, this application also provides a detection device for PE microplastics from soil mulch film, such as... Figure 4 As shown, Figure 4 This is a schematic diagram of the structure of a soil mulch film source PE microplastic detection device provided in an embodiment of this application. The soil mulch film source PE microplastic detection device 4 includes: a processor 401 and a memory 402. The memory 402 is used to store computer programs; the processor 401 is used to call and run the computer programs from the memory to execute a soil mulch film source PE microplastic detection method as described in the above embodiment.
[0079] In the embodiments of this application, the processor 401 described above can be at least one of the following: Application-Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), Central Processing Unit (CPU), Controller, Microcontroller, and Microprocessor. It is understood that for different devices, the electronic device used to implement the above processor function can also be other types, and the embodiments of this application do not specifically limit it.
[0080] This application provides a computer-readable storage medium storing a computer program for implementing, when executed by a processor, a method for detecting PE microplastics from soil mulch film sources as described in any of the above embodiments.
[0081] For example, the program instructions corresponding to the soil mulch film source PE microplastic detection method in this embodiment can be stored on storage media such as optical discs, hard disks, and USB flash drives. When the program instructions corresponding to the soil mulch film source PE microplastic detection method in the storage medium are read or executed by an electronic device, the soil mulch film source PE microplastic detection method as described in any of the above embodiments can be implemented.
[0082] Furthermore, in the embodiments of this application, the functional modules can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional module.
[0083] If the integrated unit is implemented as a software functional module and is not sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this embodiment, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the method of this embodiment. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0084] It should be understood that the phrases "one embodiment," "an embodiment," or "some embodiments" throughout the specification mean that a specific feature, structure, or characteristic related to an embodiment is included in at least one embodiment of this application. Therefore, "in one embodiment," "in one embodiment," or "in some embodiments" appearing throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It should be understood that in the various embodiments of this application, the sequence numbers of the above-described processes do not imply a sequential order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application. The sequence numbers of the above-described embodiments are merely for descriptive purposes and do not represent the superiority or inferiority of the embodiments. The descriptions of the various embodiments above tend to emphasize the differences between the various embodiments; their similarities or commonalities can be referred to mutually, and for the sake of brevity, these will not be repeated here.
[0085] The modules described above as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules. They may be located in one place or distributed across multiple network units. Some or all of the modules may be selected to achieve the purpose of this embodiment according to actual needs.
[0086] In addition, each functional module in the various embodiments of this application can be integrated into one processing unit, or each module can be a separate unit, or two or more modules can be integrated into one unit; the integrated modules can be implemented in hardware or in the form of hardware plus software functional units.
[0087] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media that can store program code, such as mobile storage devices, read-only memory (ROM), magnetic disks, or optical disks.
[0088] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.
[0089] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.
[0090] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.
[0091] The above description is merely an embodiment of this application, but the protection scope of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the protection scope of this application. Therefore, the protection scope of this application should be determined by the protection scope of the claims.
Claims
1. A method for detecting PE microplastics from soil mulch film, characterized in that, include: Acquire visible light images, raw fluorescence images, and background fluorescence images of the soil surface; Based on the visible light image of the soil surface, a preliminary screening of suspected plastic film areas is performed to obtain suspected visible light masks. Based on the original fluorescence image and the background fluorescence image, fluorescence region segmentation is performed to obtain a fluorescence region mask; Based on the visible light suspected mask and the fluorescent region mask, spatial intersection calculation and PE source discrimination of the mulch film are performed to obtain the microplastic detection results; The microplastic detection results were then converted into soil abundance.
2. The method according to claim 1, characterized in that, The process of performing preliminary screening of suspected plastic film areas based on the visible light image of the soil surface to obtain suspected visible light masks includes: The visible light image of the soil surface is preprocessed to obtain a hue channel image, a saturation channel image, and a brightness channel image; Based on the hue channel image and the saturation channel image, a coarse screening of color and brightness thresholds is performed to obtain a primary candidate mask; Based on the brightness channel image, the local standard deviation of brightness values in the neighborhood of each pixel is calculated to generate a texture contrast map; and based on the texture contrast map and a preset contrast upper limit threshold, a texture filtering mask is determined. The primary candidate mask and the texture screening mask are fused and morphologically modified to obtain the visible light potential mask.
3. The method according to claim 2, characterized in that, The preprocessing of the visible light image of the soil surface to obtain a hue channel image, a saturation channel image, and a brightness channel image includes: The visible light image of the soil surface is smoothed by a smoothing filter to obtain a smoothed visible light image; Spatial mapping is performed on the smoothed visible light image to obtain the hue channel image, the saturation channel image, and the brightness channel image.
4. The method according to claim 1, characterized in that, The step of segmenting the fluorescence region based on the original fluorescence image and the background fluorescence image to obtain a fluorescence region mask includes: The original fluorescence image is subjected to pixel-level background correction using the background fluorescence image to obtain the corrected fluorescence image; Based on the corrected fluorescence image, fluorescence region segmentation is performed to obtain the fluorescence region mask.
5. The method according to claim 4, characterized in that, The step of segmenting the fluorescence region based on the corrected fluorescence image to obtain the fluorescence region mask includes: The corrected fluorescence image is subjected to Gaussian smoothing and noise reduction processing to obtain a smoothed fluorescence image; Based on the smoothed fluorescence image, a global grayscale histogram is calculated; and a preliminary screening fluorescence mask is generated according to the global grayscale histogram and a preset optimal segmentation threshold. By setting a minimum effective fluorescence area threshold, the initial screening fluorescence mask is denoised by area thresholding to obtain the fluorescence region mask.
6. The method according to claim 1, characterized in that, The spatial intersection operation and PE source discrimination based on the visible light suspected mask and the fluorescent region mask are used to obtain microplastic detection results, including: Perform spatial intersection operation and effective target marking on the visible light suspected mask and the fluorescent region mask to obtain the spatial intersection mask and the total number of effective targets corresponding to the spatial intersection mask; For each valid target in the spatial intersection mask, joint feature extraction is performed based on the spatial intersection mask, the original fluorescence image, and the visible light image to obtain the joint features of each valid target; Calculate the comprehensive PE suspicion score for each valid target based on the joint features of each valid target; By combining the comprehensive PE suspected score of each valid target with a preset dynamic threshold, the PE source of the mulch film is identified, and the microplastic detection result is determined.
7. The method according to claim 6, characterized in that, For each valid target in the spatial intersection mask, joint feature extraction is performed based on the spatial intersection mask, the original fluorescence image, and the visible light image to obtain the joint features of each valid target, including: For each valid target in the spatial intersection mask, fluorescence features are extracted based on the original fluorescence image and the spatial intersection mask to obtain a fluorescence feature group for each valid target; Based on the spatial intersection mask and the visible light image, morphological features are extracted to obtain a morphological feature group for each effective target; A joint feature of each effective target is generated by combining the fluorescence feature set and the morphological feature set of each effective target.
8. A soil mulch film-derived PE microplastic detection system, characterized in that, include: The module comprises an acquisition module, a filtering module, a segmentation module, and a detection module; among which, The acquisition module is used to acquire visible light images of the soil surface, raw fluorescence images, and background fluorescence images; The screening module is used to perform preliminary screening of suspected plastic film areas based on the visible light image of the soil surface to obtain suspected visible light masks; The segmentation module is used to segment the fluorescence region based on the original fluorescence image and the background fluorescence image to obtain a fluorescence region mask; The detection module is used to perform spatial intersection calculation and PE source discrimination based on the visible light suspected mask and the fluorescent region mask to obtain microplastic detection results; and convert the microplastic detection results into soil abundance.
9. A soil mulch film-derived PE microplastic detection device, characterized in that, include: Processor and memory, of which, The memory is used to store computer programs; The processor is configured to call and run the computer program from the memory to perform the method as described in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that, It stores executable instructions for causing a processor to execute, thereby implementing the method of any one of claims 1 to 7.
