A method and system for detecting heavy metal micro-zone distribution of a plant processing product

Abstract

The application provides a plant processing product heavy metal micro-area distribution detection method and system, relates to the technical field of quality safety detection, and the method comprises the following steps: obtaining a cross-section micro-area section sample of a rhizome type traditional Chinese medicinal material slice, and performing frozen embedding and ultrathin section flattening and fixing treatment on the surface of the cross-section micro-area section sample to obtain a to-be-tested micro-area section; performing bright-field and dark-field composite optical microscopic imaging on the to-be-tested micro-area section to obtain an initial optical image containing plant tissue morphological characteristics, performing feature extraction on the initial optical image, and obtaining a micro-area tissue morphological mask. The application improves the heavy metal micro-area spatial distribution and tracing ability.

Description

Technical Field

[0001] This invention relates to the field of quality and safety testing technology, and in particular to a method and system for detecting the micro-regional distribution of heavy metals in processed plant products. Background Technology

[0002] Taking the heavy metal detection of Astragalus membranaceus (Huangqi) slices, a traditional Chinese medicine made from rhizomes, as an example: A quality inspection agency conducts routine heavy metal and harmful element detection on commercially available Astragalus membranaceus slices. The specific operation involves taking the crude powder of the test sample, digesting it with acid, and then performing a full element analysis using inductively coupled plasma mass spectrometry. The above-mentioned routine volume digestion detection method involves digesting the entire sample into a homogeneous solution and then measuring the total amount of elements. The result obtained is only a single concentration value representing the overall level of the entire sample. Due to the internal tissue heterogeneity of processed rhizomes, there are essential differences in morphology, chemical composition, and physiological function among different tissue regions such as the epidermis, cortex, phloem, xylem, and pith. The enrichment degree and distribution pattern of the same target element in different tissue microregions are usually quite different.

[0003] Existing volumetric digestion methods erase the aforementioned tissue-specific differences, providing only an overall average. For example, regarding the distribution of lead in Astragalus membranaceus slices, conventional detection methods cannot distinguish whether the element is concentrated in the epidermis or evenly distributed in the xylem vessels, making it difficult to answer technical questions: how is the target element distributed in different tissue micro-regions of processed plant products; whether the detected heavy metals originate from exogenous surface contamination or endogenous background enrichment. Conventional heavy metal total detection methods struggle to obtain spatial distribution information of the target element in different tissue micro-regions of processed plant products, making it difficult to distinguish whether heavy metal pollution originates from exogenous introduction or endogenous enrichment, and lacking the technical capability to trace and determine the source of pollutants. Summary of the Invention

[0004] This invention provides a method and system for detecting the micro-regional distribution of heavy metals in plant processed products, which improves the spatial distribution and traceability capabilities of heavy metals in micro-regions.

[0005] To solve the above-mentioned technical problems, the technical solution of the present invention is as follows: In a first aspect, a method for detecting the micro-regional distribution of heavy metals in processed plant products, the method comprising: Cross-sectional micro-area slices of rhizome-type Chinese medicinal materials were obtained, and the surface of the cross-sectional micro-area slices was subjected to cryo-embedding and ultrathin sectioning and fixation to obtain the micro-area slices to be tested. Bright-dark field composite optical microscopy imaging was performed on the section of the micro-area to be measured to obtain an initial optical image containing plant tissue morphology features. Feature extraction was performed on the initial optical image to obtain a micro-area tissue morphology mask. Based on the target scanning area defined by the micro-region morphology mask, laser ablation micro-region elemental surface scanning analysis is performed on the micro-region slice to be tested to obtain the two-dimensional spatial concentration distribution matrix of the target heavy metal elements on the micro-region slice to be tested. The concentration gradient feature values ​​of each plant tissue region are obtained by spatial coordinate registration and overlay mapping of the two-dimensional spatial concentration distribution matrix and the micro-region tissue morphology mask. Based on the micro-region tissue morphology mask, the geometric contour of the boundary of adjacent plant tissues is determined, and the corresponding concentration data in the two-dimensional spatial concentration distribution matrix is ​​extracted along the normal direction of the geometric contour to construct the cross-boundary concentration distribution sequence. Calculate the first-order difference between adjacent spatial nodes in the concentration distribution sequence; locate the extreme points of concentration abrupt changes by calculating the geometric rate of change of the first-order difference. By calculating the magnitude of the concentration difference within a preset neighborhood on both sides of the extreme concentration mutation point, the concentration jump rate across the boundary of adjacent plant tissues is obtained; based on the concentration gradient characteristic value and the concentration jump rate, the preset heavy metal pollution source tracing judgment rules are matched to obtain the detection results.

[0006] Secondly, a system for detecting the micro-regional distribution of heavy metals in processed plant products includes: The acquisition module is used to acquire cross-sectional micro-area slice samples of root and rhizome Chinese medicinal materials, and to perform cryo-embedding and ultrathin sectioning and fixation on the surface of the cross-sectional micro-area slice samples to obtain the micro-area slice to be tested. The extraction module is used to perform bright-dark field composite optical microscopy imaging on the micro-area slice to be tested, to obtain an initial optical image containing plant tissue morphology features, and to extract features from the initial optical image to obtain a micro-area tissue morphology mask. The analysis module is used to perform laser ablation micro-area elemental surface scanning analysis on the micro-area slice to be tested based on the target scanning area defined by the micro-area tissue morphology mask, and obtain the two-dimensional spatial concentration distribution matrix of the target heavy metal elements on the micro-area slice to be tested; The module is used to perform spatial coordinate registration and overlay mapping between the two-dimensional spatial concentration distribution matrix and the micro-region tissue morphology mask to obtain the concentration gradient feature values ​​of each plant tissue region; based on the micro-region tissue morphology mask, the geometric contour of the boundary of adjacent plant tissues is determined, and the corresponding concentration data in the two-dimensional spatial concentration distribution matrix is ​​extracted along the normal direction of the geometric contour to construct the cross-boundary concentration distribution sequence. The calculation module is used to calculate the first-order difference value of adjacent spatial nodes in the concentration distribution sequence; and to locate the extreme points of concentration abrupt changes by calculating the geometric rate of change of the first-order difference value. The matching module is used to obtain the concentration jump rate across the boundary of adjacent plant tissues by calculating the change range of the concentration difference in the preset neighborhood on both sides of the extreme value of concentration jump; based on the concentration gradient feature value and the concentration jump rate, it matches the preset heavy metal pollution source tracing judgment rules to obtain the detection results.

[0007] Thirdly, a computing device includes: One or more processors; A storage device for storing one or more programs that, when executed by one or more processors, cause the one or more processors to implement the method.

[0008] Fourthly, a computer-readable storage medium storing a program that, when executed by a processor, implements the method.

[0009] The above-described solution of the present invention has at least the following beneficial effects: By employing a synergistic integration of cryogenic embedding ultrathin sectioning, light and dark field optical microscopy to extract tissue morphology masks, mask-guided LAICPMS microarea scanning, and cross-tissue boundary concentration jump analysis techniques, this method overcomes the technical shortcomings of existing bulk digestion detection methods, which cannot obtain the spatial distribution characteristics of heavy metal elements in different tissue microareas of plant processed products and are difficult to distinguish the types of pollution sources. It achieves comprehensive technical effects such as accurate quantitative characterization of the microarea spatial distribution of target heavy metal elements in root and rhizome Chinese medicinal materials, quantitative calculation of cross-tissue boundary concentration jumps, and automatic source tracing and determination of exogenous / endogenous pollution sources. Attached Figure Description

[0010] Figure 1 This is a schematic flowchart of a method for detecting the micro-regional distribution of heavy metals in plant processed products, provided by an embodiment of the present invention.

[0011] Figure 2 This is a schematic diagram of the process for obtaining a slice of the micro-region to be tested, provided by an embodiment of the present invention.

[0012] Figure 3 This is a schematic diagram of a heavy metal micro-area distribution detection system for plant processed products provided by an embodiment of the present invention. Detailed Implementation

[0013] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0014] like Figure 1 As shown, an embodiment of the present invention proposes a method for detecting the micro-regional distribution of heavy metals in processed plant products, the method comprising the following steps: Step 1: Obtain cross-sectional micro-area slice samples of root and rhizome Chinese medicinal materials, and perform cryo-embedding and ultrathin sectioning and fixation on the surface of the cross-sectional micro-area slice samples to obtain the micro-area slice to be tested. Step 2: Perform bright-field and dark-field composite optical microscopy imaging on the micro-area slice to be tested to obtain an initial optical image containing plant tissue morphology features. Extract features from the initial optical image to obtain a micro-area tissue morphology mask. Step 3: Based on the target scanning area defined by the micro-area morphology mask, perform laser ablation micro-area elemental surface scanning analysis on the micro-area slice to be tested to obtain the two-dimensional spatial concentration distribution matrix of the target heavy metal elements on the micro-area slice to be tested. Step 4: Perform spatial coordinate registration and overlay mapping between the two-dimensional spatial concentration distribution matrix and the micro-region tissue morphology mask to obtain the concentration gradient feature values ​​of each plant tissue region; based on the micro-region tissue morphology mask, determine the geometric contour of the boundary between adjacent plant tissues, and extract the corresponding concentration data in the two-dimensional spatial concentration distribution matrix along the normal direction of the geometric contour to construct a cross-boundary concentration distribution sequence. Step 5: Calculate the first-order difference value of adjacent spatial nodes in the concentration distribution sequence; locate the extreme points of concentration abrupt change by calculating the geometric rate of change of the first-order difference value. Step 6: By calculating the change range of concentration difference in the preset neighborhood on both sides of the extreme value of concentration mutation, the concentration jump rate across the boundary of adjacent plant tissue is obtained; based on the concentration gradient feature value and the concentration jump rate, the preset heavy metal pollution source tracing judgment rule is matched to obtain the detection result.

[0015] In this embodiment of the invention, a technique combining laser ablation micro-area elemental surface scanning analysis and bright-dark field composite optical microscopic imaging is employed for collaborative detection. The two-dimensional spatial concentration distribution matrix and the micro-area tissue morphology mask are spatially registered, superimposed, mapped, and the concentration jump rate across tissue boundaries is quantitatively calculated. This overcomes the technical shortcomings of conventional bulk phase digestion detection methods, which only obtain a single average concentration after homogenizing the entire sample, cannot acquire spatial distribution information of target elements in different tissue micro-areas of plant processed products, and cannot distinguish between exogenous surface contamination and endogenous background enrichment. This achieves the technical effect of characterizing the spatial distribution of target heavy metal elements at the tissue micro-area scale, analyzing cross-boundary concentration jumps, and determining the source of pollution in root and rhizome-type Chinese medicinal herbs.

[0016] In a preferred embodiment of the present invention, such as Figure 2 As shown, step 1 above may include: Step 11: Obtain slices of rhizome-type Chinese medicinal materials and perform liquid nitrogen flash freezing on the slices to obtain cryo-solidified samples. The cryo-solidified samples are then cryo-embedded in low-temperature resin to obtain tissue embedding blocks. Specifically, this involves: taking one slice of commercially available Astragalus membranaceus, visually inspecting its appearance, and selecting a region in the center of the slice that is free of cracks, obvious mechanical damage, and has a clear cross-sectional texture. From this selected region, cut a circular cross-sectional slice approximately 3 mm thick and 15 mm in diameter perpendicular to the longitudinal axis of the slice to obtain a slice of rhizome-type Chinese medicinal materials. Based on this slice, perform liquid nitrogen flash freezing. The specific procedure is as follows: Slices of root and rhizome-type Chinese medicinal herbs are placed in a stainless steel sample basket and immersed in a Dewar flask containing liquid nitrogen at -196 degrees Celsius for 45 seconds. During this time, free water within the tissue rapidly forms tiny ice crystals with a particle size ranging from nanometers to submicrometers under the instantaneous low temperature, inhibiting the mechanical compression and damage to the ultrastructure of cells caused by the growth of large ice crystals. After the rapid freezing is complete, the sample is removed, resulting in a frozen-solidified sample. This frozen-solidified sample is in a firm state, and the water within the tissue has been fixed by ice crystals.

[0017] Based on the cryo-cured samples, cryo-embedding was performed using low-temperature resin. The specific procedure was as follows: In a cryogenic chamber maintained at -25°C, the cryo-cured sample was transferred to a pre-cooled silicone embedding mold with internal dimensions of 20mm × 10mm × 5mm. Using a pipette, pre-cooled London white acrylic resin (pre-cooled to -25°C) was slowly injected along the inner wall of the mold, gradually raising the liquid level until the sample was submerged, avoiding air bubbles during injection. The completed embedding mold was then transferred to a -25°C UV polymerization chamber, using a 365nm UV-emitting diode array as the irradiance source, with a UV irradiance of [missing information]. After continuous irradiation for 48 hours, the acrylic resin monomers undergo free radical polymerization under the combined action of ultraviolet light and low temperature, gradually transforming from a liquid state into a solid cross-linked network structure. After polymerization is completed, the solid resin block is removed from the mold to obtain a tissue embedding block. Inside the tissue embedding block, the Astragalus membranaceus slices are embedded in the resin and firmly supported, with the cross-sectional direction of the slices parallel to the top surface of the embedding block.

[0018] Step 12: The tissue embedding block is transversely cut using an ultramicrotome to obtain cross-sectional microsection samples. Specifically, the tissue embedding block is clamped in the sample arm chuck of the Leica EMUC7 ultramicrotome. The sample arm posture is adjusted so that the top surface of the tissue embedding block is parallel to the plane of the diamond scalpel cutting edge. The diamond scalpel is installed with a cutting angle of 45 degrees. Deionized water is injected into the scalpel groove as a floating medium for the sections. The section thickness is set to 25 micrometers, the sample arm feed speed is 1.0 mm / s, and the cutting stroke is 2.0 mm. The automatic sectioning cycle program is started. The sample arm moves the tissue embedding block downwards at the set feed speed. The diamond scalpel cuts the surface of the resin block along a direction parallel to the cross-section of the Astragalus membranaceus slices, with each feed being 25 micrometers. The cut sections then float on the water surface in the scalpel groove, forming a series of sections with uniform thickness. Using the tip of a fine brush, pick up intact, uncurled, and untorn sections from the water surface and transfer them to a standard microscope slide measuring 76mm × 26mm × 1mm. Allow them to air dry at room temperature for 10 minutes to ensure close adhesion between the sections and the slide surface. After drying, a cross-sectional microsection sample is obtained. This cross-sectional microsection sample completely contains all tissue layers from the epidermis to the medulla. Under a 20x optical microscope, the epidermis (appearing as a narrow dark ring), cortex (appearing as a wider light-colored area), phloem (appearing as discontinuous dark rings), xylem (appearing as fan-shaped radial vessels), and medulla (located in the central region) can be clearly distinguished.

[0019] Step 13 involves surface smoothing and morphology fixation of the cross-sectional micro-area slice sample to eliminate microscopic morphological fluctuations on the slice surface, resulting in the micro-area slice to be tested. Specifically, this includes: performing surface smoothing and morphology fixation on the cross-sectional micro-area slice sample to eliminate microscopic morphological fluctuations on the slice surface. During the preparation process, the slice surface may develop microscopic wrinkles or unevenness with height fluctuations ranging from several micrometers to tens of micrometers due to factors such as blade cutting force, resin shrinkage stress, and uneven moisture evaporation during drying. Since laser ablation analysis requires that the height variation of the sample surface not exceed the focal depth of the laser beam, surface smoothing is performed on the slice surface. The specific operation is as follows: placing the glass slide containing the slice in a vacuum drying oven and continuously evacuating for 30 minutes at a temperature of 25 degrees Celsius and a vacuum degree of -0.08 MPa to remove residual trace moisture from the slice tissue and resin matrix. The glass slide is then transferred to a precision hot-pressing flattening device, which consists of an upper plate heating module, a lower stage, and a pressure sensor.

[0020] The upper pressure plate temperature is set to 60 degrees Celsius, and a uniform pressure of 0.5 MPa is applied through a precision pressure regulating valve. The hot pressing is continued for 5 minutes. Under the combined effects of temperature and pressure, the microscopic protrusions on the surface of the slice are flattened, and the resin matrix undergoes a small amount of thermal softening and filling, resulting in a smoother surface morphology. After the hot pressing is completed, the heating power is turned off, and the device is allowed to cool naturally to room temperature of 25 degrees Celsius under pressure for about 15 minutes. The pressure is then removed, the glass slide is taken out, and the micro-area slice to be tested is obtained. The root mean square value of the surface height undulation of the micro-area slice to be tested does not exceed 2 micrometers, which meets the requirements of laser ablation for the surface smoothness of the sample. At the same time, the temperature during the hot pressing process does not exceed the glass transition temperature of the resin, so the tissue morphology and structure are completely preserved.

[0021] In this embodiment of the invention, a sample preprocessing technique combining liquid nitrogen quick-freezing and low-temperature resin cryopreservation with ultrathin sectioning and planarization is employed. This is combined with bright-and-dark field composite optical microscopy for multi-channel tissue morphology feature extraction and micro-area tissue morphology mask construction. The mask is used to define the target area for laser ablation micro-area elemental surface scanning analysis. The two-dimensional spatial concentration distribution matrix and the micro-area tissue morphology mask are spatially registered, superimposed, mapped, and cross-boundary normal direction concentration sequences are extracted. The first-order differential geometric change rate is used to locate extreme concentration abrupt points and calculate the concentration sequence across adjacent plant tissue boundaries. The concentration jump rate at the boundary overcomes the technical shortcomings of conventional bulk digestion detection methods, which can only obtain a single average concentration after homogenizing and digesting the whole sample. These methods cannot obtain spatial distribution information of target heavy metal elements in different tissue micro-regions of plant processed products, cannot distinguish between exogenous surface pollution and endogenous background enrichment, and are difficult to achieve source tracing of pollution at the tissue micro-region scale. The method achieves comprehensive technical effects in characterizing the spatial distribution of target heavy metal elements at the tissue micro-region scale, quantitatively analyzing concentration jumps across tissue boundaries, and accurately tracing the source of exogenous / endogenous pollution types in root and rhizome Chinese medicinal materials.

[0022] In a preferred embodiment of the present invention, step 2 above may include: Step 21: Multi-channel imaging acquisition of the micro-area section to be tested using a bright-field and dark-field combined optical microscope to obtain an initial optical image containing plant tissue morphological characteristics; grayscale conversion and contrast enhancement processing of the initial optical image to obtain an enhanced grayscale image, specifically including: multi-channel imaging acquisition of the micro-area section to be tested using a bright-field and dark-field combined optical microscope, specifically: placing the micro-area section to be tested on the motorized stage of the bright-field and dark-field combined optical microscope, and fixing both ends of the slide with spring clips. First, operate in bright-field imaging mode, with a halogen lamp as the transmitted illumination source, the color temperature of the light source set to 3200 Kelvin, and uniformly illuminating the section using Kohler illumination after adjustment of the condenser lens and aperture diaphragm. A 20x plan achromatic objective lens with a numerical aperture of 0.50 was selected. The microscope's motorized stage automatically scanned along the row and column directions using a serpentine path driven by the control software. The overlap rate between adjacent fields of view was set to 25%. The images of each field of view acquired by the scan were automatically fused by the image stitching algorithm built into the control software to generate a panoramic image covering the entire cross-sectional area of ​​the slice. The stitching algorithm adopted a phase-correlation-based translation estimation method, and the stitching accuracy reached the sub-pixel level.

[0023] After the bright-field channel scan is completed, the microscope is switched to dark-field imaging mode. The inner numerical aperture of the annular dark-field condenser is 0.80 to 0.95, allowing direct illumination light to pass outside the objective aperture angle without entering the objective lens. Only the scattered light generated by the tissue section on the incident light is collected and imaged by the objective lens. In dark-field mode, a second full-area scan is performed using the same 20x objective lens and the same scanning path to obtain a panoramic image of the dark-field channel. The two panoramic images of bright-field and dark-field are naturally pixel-aligned in space. The two panoramic images are stored as 16-bit tagged image files to obtain an initial optical image containing the morphological features of the plant tissue. In the initial optical image, the bright-field channel clearly presents the cell wall contours, intercellular spaces, and overall tissue hierarchy of each tissue region. The dark-field channel, through the contrast of scattered light, highlights the pits on the cell walls, the contours around starch grains, and the secondary thickening structures on the vascular bundle vessel walls, which have strong light scattering characteristics. Based on this initial optical image, grayscale and contrast enhancement processing is performed. The specific operation of grayscale processing is as follows: The color panoramic images of the bright field channel and the dark field channel are both red, green and blue three-channel color images. The red component pixel value, green component pixel value and blue component pixel value of each channel are extracted respectively, and then synthesized into a single-channel grayscale image according to the following weighted grayscale conversion formula: ; In the formula, This represents the pixel value of the synthesized grayscale image, with a value range of 0 to 65535 (16-bit image). This represents the pixel value of the red channel, ranging from 0 to 65535; This represents the pixel value of the green channel, ranging from 0 to 65535; This represents the pixel value of the blue channel, ranging from 0 to 65535; , , These represent the weighting coefficients for the red, green, and blue channels corresponding to the relative luminance function in the International Commission on Illumination (ICI) standard, respectively. All are dimensionless constants. After performing the above grayscale processing on the bright field and dark field channels, bright field grayscale images and dark field grayscale images are obtained. The two grayscale images are then superimposed with equal weights using the following formula: ; In the formula, Represents the pixel coordinates of a composite grayscale image. Pixel value at; Indicates a bright-field grayscale image in Pixel value at; Indicates a dark-field grayscale image in Pixel value at; The equal-weighted superposition coefficients are dimensionless. Contrast-limited adaptive histogram equalization is performed on the composite grayscale image, which uniformly divides the composite grayscale image into... For each local block, a grayscale histogram is calculated, and the number of pixels exceeding the clipping limit of 0.01 is evenly redistributed across the grayscale levels of the histogram. Histogram equalization is then performed on each block. Finally, bilinear interpolation is applied to pixels at the boundaries of adjacent blocks to eliminate artifacts between blocks. The resulting enhanced grayscale image improves the boundary contrast between tissue regions, increasing the grayscale differences between the epidermis and cortex, and between the phloem and xylem, providing clear grayscale transition features for subsequent edge detection.

[0024] Step 22 involves performing edge detection and morphological closing operations on the enhanced grayscale image to extract the regional boundary contours of the epidermis, cortex, phloem, and xylem, obtaining an initial boundary contour map. Specifically, this includes applying a two-dimensional Gaussian filter kernel to the enhanced grayscale image, with a kernel size of [size missing]. Gaussian function standard deviation A value of 1.4 pixels was chosen to suppress high-frequency noise components in the image while maintaining the sharpness of tissue boundaries. For each pixel in the filtered image, horizontal and vertical grading was applied respectively. The Sobel operator template is used to compute gradient components. The template row vectors of the horizontal Sobel operator are as follows: (First line) (Second line) (Third row), the template row vectors of the vertical Sobel operator are as follows: (First line) (Second line) (Third row) Pixel position coordinates gradient magnitude at The calculation formula is: ; In the formula, For pixels in The gradient magnitude at that point has the same dimensions as the pixel grayscale value. The horizontal gradient component is obtained by convolving a horizontal Sobel template with a local region of the image. The vertical gradient component is obtained by convolving the vertical Sobel template with a local region of the image, and the gradient direction is... The calculation formula is: ; In the formula, Indicates the pixel in The gradient direction at a given point, in radians, has a range of values ​​of [value missing]. to ; This represents the arctangent function. Along the gradient direction of each pixel, it compares the gradient magnitude of that pixel with the gradient magnitude of its two adjacent pixels in that direction. If the gradient magnitude of the pixel is not the maximum of the three, its gradient magnitude is set to zero; otherwise, it is retained. After this processing, the edge response is compressed into a ridge with a width of one pixel. The preset low threshold is 25, and the high threshold is 75. Pixels with a gradient magnitude greater than 75 are directly marked as strong edge pixels. Pixels with a gradient magnitude between 25 and 75 are marked as weak edge pixels and retained if there is a marked strong edge pixel in their eight neighborhoods; otherwise, they are discarded. Pixels with a gradient magnitude less than 25 are directly discarded. After processing, an initial binary edge map is obtained, where edge pixels have a value of 1 and background pixels have a value of 0. Based on the initial binary edge map, morphological closing operations are performed.

[0025] The morphological closing operation is performed by first dilating and then eroding. Both operations use a disk-shaped structuring element with a radius of 5 pixels. The dilution operation slides the structuring element pixel by pixel on the edge binary image, taking the maximum pixel value within the area covered by the structuring element as the output, connecting adjacent edge segments with minor breaks. The erosion operation follows immediately, taking the minimum pixel value within the covered area of ​​the same structuring element as the output, shrinking the expanded edge contour back to a position approximately the original width. The result of the closing operation is that the edge breaks caused by insufficient local contrast in the initial edge binary image are effectively closed, but the true spatial position of the tissue boundary is not changed. The edge binary image after the closing operation is superimposed with an enhanced grayscale image for visual verification, confirming that the outer boundary of the epidermis, the boundary between the epidermis and the cortex, the boundary between the cortex and the phloem, the boundary between the phloem and the xylem, and the boundary between the xylem and the pith are all extracted continuously and accurately, resulting in the initial boundary contour image.

[0026] Step 23: Perform connected component geometric segmentation and region filling processing on the initial boundary contour map to obtain a micro-area tissue morphology mask containing spatial location information of each tissue region. Specifically, this includes: performing connected component geometric segmentation processing based on the initial boundary contour map. The specific operation is as follows: the closed boundary curve in the initial boundary contour map divides the entire image plane into multiple unconnected closed regions. Starting from the four corner pixels of the image, the four-neighbor flooding filling algorithm is used to mark all pixels connected to the image background as background. For each of the remaining closed regions that are not marked as background, a unique positive integer label value is sequentially detected and assigned. The label values ​​are assigned based on the spatial location characteristics of the regions in the cross-section, determined sequentially from the outside in: Label value 1 corresponds to the outermost narrow, closed ring region, the epidermis; Label value 2 corresponds to the wider region inside the epidermis, the cortex; Label value 3 corresponds to the discontinuous ring region inside the cortex containing sieve tubes and companion cells, the phloem; Label value 4 corresponds to the region inside the phloem with the largest area, arranged in a fan-shaped radial pattern, and containing vessel elements, the xylem; Label value 5 corresponds to the central circular region, the pith. Based on the connected component label assignment results, region filling is performed. For each closed connected component with assigned label values, all pixels within that component are traversed, and the grayscale value of each pixel is set to the same label value as the boundary of that connected component. Pixels with the same label value constitute a complete tissue region, and the spatial distribution of label values ​​constitutes a label image with the same pixel size as the enhanced grayscale image. The labeled image is a micro-region tissue morphology mask. The micro-region tissue morphology mask uses the image coordinate system as a reference. The label value of each pixel identifies the tissue category to which the pixel belongs, and the row and column coordinates of the pixel identify its spatial location.

[0027] In this embodiment of the invention, by employing the technique of multi-channel imaging acquisition using bright and dark field composite optical microscopy, the initial optical image is sequentially subjected to grayscale conversion and contrast enhancement, edge detection and morphological closing operation, connected domain geometric segmentation and region filling processing. This overcomes the technical defects of incomplete acquisition of morphological information of different structural layers of plant tissues, blurred tissue region boundaries making accurate segmentation difficult, and inability to provide precise target region definition for subsequent micro-area elemental surface scanning analysis under single bright field or dark field imaging modes. It achieves the technical effect of accurately extracting the boundary contours of various plant tissue regions such as epidermis, cortex, phloem and xylem and obtaining a high-precision micro-area tissue morphology mask containing the spatial location information of each tissue region.

[0028] In a preferred embodiment of the present invention, step 3 above may include: Step 31: Extract the spatial coordinate range of each plant tissue region in the micro-region tissue morphology mask to obtain the target scanning area. Specifically, this includes: extracting the spatial coordinate range of each plant tissue region based on the micro-region tissue morphology mask. The specific operation is as follows: traverse all pixels in the micro-region tissue morphology mask one by one, and for each of the five connected components corresponding to label values ​​1 to 5, record the minimum value of the row coordinates of all pixels in each connected component. and maximum value and the minimum value of the column coordinates. and maximum value ,in The label values, taking values ​​of 1, 2, 3, 4, and 5, form a bounding box whose coordinate range defines the smallest bounding rectangle corresponding to each tissue region in the image coordinate system. The pixel coordinate values ​​of the bounding boxes are multiplied by a calibration coefficient to convert them into actual physical coordinates, thus obtaining the target scanning area corresponding to each tissue region. Taking the epidermal region as an example, its bounding box and the corresponding target scanning area define the physical range of the laser ablation scan within the epidermal region. The target scanning area of ​​the xylem region covers a fan-shaped area from inside the cambium ring to outside the pith boundary. The target scanning area defines the spatial boundaries of each rectangular sub-region in which the laser ablation device needs to perform point-by-point ablation operations in its working coordinate system using actual physical coordinates.

[0029] Step 32: Based on the target scanning area, control the laser ablation device to perform point-by-point array ablation on the micro-area slice to be tested, obtaining an aerosol sample flow. Specifically, this includes: according to the target scanning area, controlling the laser ablation device to perform point-by-point array ablation on the micro-area slice to be tested. The operating parameters of the laser ablation device are set as follows: laser output wavelength 213 nm, single pulse energy 3.5 mJ, pulse repetition frequency 10 Hz, ablation spot diameter 30 μm, and laser energy density on the sample surface approximately [value missing]. The carrier gas is high-purity helium, with a flow rate of [missing information]. The specific scanning path is planned as follows: For each rectangular sub-region in the target scanning area, starting from the upper left corner coordinate point of the sub-region, laser pulses are emitted point by point along the row direction with a fixed step size of 30 micrometers to perform ablation. After completing the scanning of all points in the row, the laser pulses are shifted 30 micrometers along the column direction and the next row is scanned in reverse, forming a serpentine reciprocating scanning trajectory. The row spacing is set to 30 micrometers to ensure that adjacent ablation points are closely adjacent in space without any gaps, achieving seamless full-coverage scanning of the entire target area of ​​the micro-slice to be tested. At each ablation point, a single laser pulse locally and instantaneously heats the sample surface to above the vaporization temperature within a circular area with a diameter of 30 micrometers. The generated aerosol particles are carried away from the sample surface by the helium carrier gas flowing through the sample cell and transported to the downstream detection instrument along the transmission pipeline. During the laser point-by-point ablation process, each ablation point is ablated in chronological order, and the corresponding aerosol particles enter the transmission pipeline in the same chronological order, forming an aerosol sample stream arranged along the time axis. Each time segment in this aerosol sample stream corresponds to a sample particle at an ablation point, maintaining a strict correspondence between the spatial position of the point and the time sequence.

[0030] Step 33: Elemental isotope signals of the aerosol sample stream are acquired using a mass spectrometer to obtain the target heavy metal element signal intensity values ​​at each ablation point. Specifically, this includes: Based on the aerosol sample stream, elemental isotope signals are acquired using a mass spectrometer. After leaving the sample cell of the laser ablation device, the aerosol sample stream passes through a 4 mm inner diameter PTFE transmission pipe and is mixed online with an argon supplementary gas flow rate of 0.9 liters per minute with a purity of not less than 99.996%. The mixed gas flow enters the inductively coupled plasma (ICP) tube. The operating parameters of the ICP are: RF generator forward power 1550 watts, reverse... The plasma power is less than 5 watts; the plasma cooling gas is argon with a flow rate of 14.0 liters per minute; the auxiliary gas is argon with a flow rate of 0.8 liters per minute; the nebulizing gas is argon with a flow rate of 1.05 liters per minute. Under the action of high-temperature plasma of 6000 to 8000 Kelvin, aerosol particles undergo desolution, evaporation, atomization, and ionization processes. Lead and other matrix elements in the sample are ionized into positively charged single-charged ions. After the ion beam is extracted by two stages of differential pressure in the interface region through a sampling cone with an aperture of 1.0 mm and a cutoff cone with an aperture of 0.7 mm, it enters the collision reaction cell and the triple quadrupole mass analyzer. High-purity helium is introduced into the collision reaction cell at a flow rate of 0.25 L / min to eliminate the interference of oxides and polyatomic ions on the detection of target elements.

[0031] The quality analyzer operates in peak-skipping scan mode, selectively monitoring the following quality number channels: quality number 208 channels, corresponding to 208Pb isotopes, with a natural abundance of 52.4%, are used as quantitative isotopes for the target heavy metal element; mass number 13 channels, corresponding to 13 The carbon isotope, with a natural abundance of 1.07%, is used as an internal standard to correct signal fluctuations caused by instrument sensitivity drift and matrix effects. The dwell time for each mass number channel is 30 milliseconds. The signal acquisition period for a single ablation site is synchronized with the laser pulse period, at 100 milliseconds, corresponding to a pulse repetition frequency of 10 Hz. Within the 100-millisecond acquisition period for each ablation site, the signal generated by one laser pulse is acquired and time-resolved integration is performed. After acquisition is completed at each ablation site, the signal intensity value of the target heavy metal element at that site is output. 208 The Pb signal strength, measured in counts per second, is the same as the signal strength value of the internal standard element. 13 C signal strength, measured in counts per second. After all erosion points have been collected, a pair of signal strength values ​​are obtained for each point, with their row index... i and column indexes j Corresponding to the array position of the laser ablation scan.

[0032] Step 34: Quantitatively convert the target heavy metal element signal intensity values ​​at each erosion point using a pre-set standard material calibration curve to obtain the two-dimensional spatial concentration distribution matrix of the target heavy metal element on the micro-area slice to be tested. Specifically, this includes: quantitatively converting the target heavy metal element signal intensity values ​​at each erosion point using a pre-set standard material calibration curve. The calibration curve is established by inserting three laser ablation line scans of the standard reference material at the beginning and end of the formal sample analysis sequence, for a total of six analyses. The standard reference material used is NIST SRM612, and the laser ablation parameters are consistent with the sample analysis parameters. The concentration distribution matrix of the target heavy metal element on the micro-area slice to be tested is obtained from this process. and The signal strength was determined by taking the arithmetic mean of the results of six standard substance analyses. Average signal strength and The quantitative conversion of average signal intensity and lead concentration at each erosion site was performed using the relative sensitivity factor method based on internal standard normalization. The calculation formula is as follows: ; In the formula, This indicates the row index of the erosion point in the scan array. Column index The mass fraction of lead at a given point is expressed in micrograms per gram, with the dimension being mass / mass, i.e., a dimensionless ratio. This indicates the location of the erosion. Signal strength, measured in counts per second; This indicates the location of the erosion. Signal strength, measured in counts per second; This indicates the certified mass fraction of lead in the standard reference NIST RM612, with a value of 38.57 micrograms per gram. Indicating 6 standard substance analyses The average signal strength, expressed in counts per second; Indicating 6 standard substance analyses The average signal strength, measured in counts per second, is given for any given erosion point. and Substituting the signal strength value into the above formula and combining it with the analysis of standard materials yields the results. , and The certification value is used to calculate the lead mass fraction at that location. The above calculations were performed point-by-point on all erosion sites. The calculation results for each site were arranged according to their spatial scanning positions, forming a two-dimensional numerical matrix with the number of rows equal to the total number of rows and columns of the scan. The matrix elements correspond to spatial positions, and each matrix element stores the lead mass fraction at its corresponding spatial coordinates. The concentration sub-matrices of different tissue regions were spliced ​​and merged in a common physical coordinate system to obtain a two-dimensional spatial concentration distribution matrix of the target heavy metal element on the micro-area slice to be tested. This matrix fully reflects the mass fraction distribution of lead at various spatial positions across the entire cross-section of the micro-area slice to be tested.

[0033] In this embodiment of the invention, a technique is employed to extract the spatial coordinate range of each plant tissue region based on a micro-region tissue morphology mask to limit the target scanning area. This is combined with a laser ablation device to perform point-by-point array ablation on the micro-region slice to be tested, and a mass spectrometry instrument to collect elemental isotope signals from the aerosol sample stream. Finally, the signal intensity value is quantitatively converted using a preset standard material calibration curve. Therefore, this method overcomes the technical defects of conventional whole digestion detection methods, which cannot obtain the spatial distribution information of target heavy metal elements in different tissue micro-regions of plant processed products, traditional single-point or random line scan analysis lacks tissue morphology spatial constraints, resulting in the inability to accurately correspond the detection points to the tissue regions, and the inability to accurately convert signal intensity values ​​into concentration data with spatial coordinate attributes. This method achieves the technical effect of precise positioning and ablation of target heavy metal elements in micro-region surface scanning under tissue morphology spatial constraints, efficient acquisition and quantitative calibration of elemental isotope signals at each ablation point, and finally obtaining a two-dimensional spatial concentration distribution matrix of target heavy metal elements with spatial location attributes on the micro-region slice to be tested.

[0034] In a preferred embodiment of the present invention, step 4 above may include: Step 41 involves spatial coordinate registration and overlay mapping of the two-dimensional spatial concentration distribution matrix and the micro-area tissue morphology mask to obtain a spatially aligned composite mapping matrix. Specifically, the micro-area tissue morphology mask is derived from an optical microscopic imaging system, with its spatial reference being the image coordinate system and a spatial resolution of 2.5 micrometers per pixel. This resolution is determined by the 20x objective lens and the camera pixel size. The two-dimensional spatial concentration distribution matrix is ​​derived from a laser ablation mass spectrometry scanning system, with its spatial reference being the physical working coordinate system of the laser ablation device and a spatial resolution of 30 micrometers per pixel, equal to the laser ablation spot diameter and scanning step size. The origins, coordinate axis directions, and spatial resolutions of the two coordinate systems are different, and they are unified through spatial transformation. The specific operation of spatial coordinate registration is as follows: Extract several feature anchor points of the outer contour of the cross-section of the micro-area slice to be measured from the two-dimensional spatial concentration distribution matrix, including the uppermost, lowermost, leftmost, and rightmost endpoints of the cross-section. Their physical coordinates are determined by the position where the concentration value in the concentration matrix jumps from zero to a positive value. Extract the image coordinates of feature anchor points corresponding to the same anatomical position from the micro-area tissue morphology mask. Based on at least four pairs of feature anchor points, calculate the affine transformation parameters. The mathematical expression for the affine transformation is: ; ; In the formula, Represents the column coordinates of pixels in the micro-region morphology mask, in pixels; Represents the row coordinates of pixels in the micro-region morphology mask; This represents the column coordinates in the physical coordinate system of the concentration distribution matrix after affine transformation, in micrometers; This represents the row coordinates in the physical coordinate system of the concentration distribution matrix after affine transformation, in micrometers; The scaling and rotation parameter in the row direction is a scaling factor in the column direction from the image coordinates to the physical coordinates. It is determined by the ratio of the spatial resolution of the two imaging modes and is expressed in micrometers per pixel. The shearing parameter representing the row direction is set to 0 in this embodiment because the axes of the two coordinate systems are aligned. The shearing parameter represents the column direction; in this embodiment, the value is 0. The scaling and rotation parameter in the column direction is a scaling factor in the row direction from the image coordinates to the physical coordinates, and the unit is micrometers per pixel. This represents the translation parameter along the column direction, in micrometers. The translation parameter represents the translation along the row direction, in micrometers. The value of the translation parameter is chosen to make the geometric center of the slice coincide in both coordinate systems.

[0035] Using the aforementioned affine transformation parameters, coordinate transformation is performed on each pixel coordinate of the micro-area tissue morphology mask, mapping the tissue category label information in the mask from the image coordinate system to the physical coordinate system used by the concentration distribution matrix. After the transformation, for each spatial node in the two-dimensional spatial concentration distribution matrix, the tissue region to which the node belongs in the transformed mask is determined by coordinate lookup. The tissue category label is then associated and stored with the lead element mass fraction of the node, resulting in a spatially aligned composite mapping matrix. Each data record in this composite mapping matrix contains four data fields: the first field is the row coordinate index, the second field is the column coordinate index, the third field is the lead element mass fraction value at that coordinate (in micrograms per gram), and the fourth field is the tissue region category label to which that coordinate belongs, with a value ranging from 1 to 5. This composite mapping matrix achieves pixel-level fusion of tissue morphology information and elemental concentration chemical information in a unified spatial coordinate system.

[0036] Step 42: Based on the composite mapping matrix, calculate the concentration gradient feature values ​​of the target heavy metal element in each plant tissue region; extract the geometric contours of adjacent plant tissue boundaries in the micro-region tissue morphology mask to obtain boundary contour lines. Specifically, the concentration gradient feature values ​​are used to quantify the drastic and directional changes in the lead element mass fraction within each tissue region in space. For label values... The organizational region is defined by extracting the set of all spatial nodes within that region from the composite mapping matrix, denoted as […]. Set size Equal to the total number of spatial nodes in the region, for For each spatial node, calculate its partial concentration derivatives along the row and column directions using a central difference scheme, and include the concentration gradient components along the row direction. The calculation formula is: ; In the formula, Indicates spatial location The concentration gradient component along the direction of travel, in micrograms per gram per micrometer; This indicates the mass fraction of lead in the preceding adjacent node along the row direction, expressed in micrograms per gram. It represents the mass fraction of lead in the next adjacent node along the row direction, in micrograms per gram; This represents the spatial step size along the row direction, taken as 15 micrometers, used to calculate the concentration gradient component along the column direction. Its spatial step size The value is also set to 15 micrometers. The concentration gradient amplitude at this spatial node. The calculation formula is: ; In the formula, Indicates spatial location The concentration gradient amplitude at a given point, expressed in micrograms per gram per micrometer; This represents the concentration gradient component along the column direction, in micrograms per gram per micrometer. For the entire tissue region... Its concentration gradient characteristic value Defined as the arithmetic mean of the concentration gradient magnitudes of all spatial nodes within the region, the calculation formula is: ; In the formula, Indicates the organization area The concentration gradient characteristic value, in micrograms per gram per micrometer; Indicates the organization area The total number of spatial nodes within the area is dimensionless; calculate the concentration gradient characteristic values ​​of each tissue region corresponding to label values ​​1 to 5 sequentially according to the above formula. to This yields a set of quinary concentration gradient eigenvalues. The larger the value, the more drastic the spatial variation of the lead mass fraction within the tissue region, and the more uneven the concentration distribution. The smaller the value, the more uniform the concentration distribution within the region. The geometric contours of adjacent plant tissue boundaries are extracted from the micro-region tissue morphology mask. Specifically, the operation involves detecting pixel locations where the label value changes pixel-by-pixel in the micro-region tissue morphology mask. When the label values ​​of two adjacent pixels are different, the interface between these two pixels is the tissue boundary. All detected boundary pixels are connected into closed ordered point sequences according to their spatial adjacency relationships, with each ordered point sequence corresponding to a tissue boundary contour line. Four boundary contour lines are extracted sequentially: the first is the boundary contour line between the epidermis and cortex, i.e., the pixel intersection line between label value 1 and label value 2, denoted as the boundary. The second line is the boundary outline between the cortex and phloem, that is, the pixel boundary line between label value 2 and label value 3, denoted as the boundary. The third line is the boundary outline between the phloem and xylem, that is, the pixel intersection line of label value 3 and label value 4, denoted as the boundary. The fourth line is the boundary outline between the xylem and pith, that is, the pixel intersection line of label value 4 and label value 5, denoted as the boundary. Each boundary contour line is stored as an ordered list of sample point coordinates in the physical coordinate system after affine transformation, and the arc length interval between adjacent sample points is normalized to 15 micrometers by linear interpolation.

[0037] Step 43: Along the normal direction of the boundary contour line, extract the corresponding concentration data from the two-dimensional spatial concentration distribution matrix to construct a cross-boundary concentration distribution sequence. Specifically, this includes: taking the boundary between the epidermis and dermis as an example, determining a set of sampling base points along the boundary contour line of the epidermis and dermis at equal arc length intervals. The arc length interval is 15 micrometers, and the number of sampling base points depends on the total arc length of the boundary contour line. For each sampling base point, calculate the local tangent direction vector of the boundary contour line at that base point. The specific method is as follows: take one adjacent sampling point before and after the sampling base point along the boundary contour line, and record them as the previous adjacent point and the next adjacent point, respectively. Construct two direction vectors with the sampling base point itself as the origin: the forward vector points from the base point to the next adjacent point, and the backward vector points from the base point to the previous adjacent point. The local tangent direction is taken as the normalized average of the above two direction vectors. Rotate the local tangent direction vector counterclockwise by 90 degrees to obtain the normal direction vector. The positive normal direction is defined as the direction from the epidermis to the interior of the cortex, that is, from the low label value to the high label value, and the negative normal direction is the opposite direction.

[0038] Using the sampling base point as the origin, a concentration sampling line segment containing 11 spatial nodes is obtained by extending 75 micrometers along both the positive and negative normal directions, spanning 5 spatial nodes each, with a node spacing of 15 micrometers. The 5 nodes at the negative normal end are numbered 1 to 5 and located on the epidermal side; the origin node is numbered 6 and located at the boundary line; and the 5 nodes at the positive normal end are numbered 7 to 11 and located on the cortical side. From the interpolated two-dimensional spatial concentration distribution matrix, the corresponding lead element mass fraction values ​​are extracted according to the physical coordinates of each spatial node, forming a cross-boundary concentration distribution subsequence of length 11 at the sampling base point. All concentration distribution subsequences generated by the sampling base points on the epidermal-cortical boundary are aligned according to the spatial node numbers 1 to 11, and the arithmetic mean of the concentration values ​​of all subsequences at each spatial node number is calculated to generate an averaged cross-boundary concentration distribution sequence.

[0039] The first of the sequence m There are elements, where the element index is 1. m The value ranges from 1 to 11, representing the th value along the normal direction at that boundary. m The average lead mass fraction at each spatial node position, this sequence represents the spatial variation trend of lead from the negative normal end (inside the epidermis) to the positive normal end (inside the cortex), quantitatively reflecting the concentration decay or jump characteristics of lead as it crosses the boundary between the epidermis and cortex. Following the above process, cross-boundary concentration distribution sequences were constructed for the cortex-phloem boundary, the phloem-xylem boundary, and the xylem-pith boundary. Each sequence has 11 spatial nodes, and each sequence corresponds to an averaged concentration variation profile along the normal direction at the corresponding tissue boundary.

[0040] In this embodiment of the invention, a composite mapping matrix with spatial alignment is obtained by registering and overlaying a two-dimensional spatial concentration distribution matrix with a micro-region tissue morphology mask. Based on this, the concentration gradient feature values ​​in each plant tissue region are calculated and the geometric contour lines of adjacent tissue boundaries are extracted. Furthermore, concentration data is extracted along the normal direction of the boundary contour lines to construct a cross-boundary concentration distribution sequence. Therefore, this invention overcomes the technical defects caused by the rotation and translation distortion between the laser ablation spatial coordinate system and the optical microscopic image pixel coordinate system, which makes it impossible to accurately correspond the element concentration data with the tissue morphology position. Conventional overall analysis methods are difficult to obtain information on the concentration change trend across tissue boundaries, and it is difficult to distinguish whether the elements are gradually diffused or abruptly enriched in different tissues from the perspective of spatial distribution. This invention achieves the technical effect of pixel-level accurate registration of element spatial concentration distribution and plant tissue morphology structure under a unified coordinate framework, quantitative characterization of concentration gradient features in each tissue region, and construction of a high-resolution cross-boundary concentration distribution sequence along the normal direction of adjacent tissue boundaries.

[0041] In a preferred embodiment of the present invention, step 5 above may include: Step 51: Calculate the concentration difference between adjacent spatial nodes in the cross-boundary concentration distribution sequence to obtain a set of first-order difference values. Specifically, this includes: calculating the first-order difference values ​​between adjacent spatial nodes in each cross-boundary concentration distribution sequence. The first-order difference value characterizes the point-by-point change in concentration between adjacent nodes along the cross-boundary direction. For a given cross-boundary concentration distribution sequence, its sequence length is denoted by the total number of nodes, typically 11 nodes. The formula for calculating the first-order difference value is: ; In the formula, Indicates the first The first-order difference value of each pair of adjacent nodes, in micrograms per gram, with positive values ​​indicating increasing concentration and negative values ​​indicating decreasing concentration; This indicates the first [number] in the cross-boundary concentration distribution sequence. Concentration values ​​at each node, in micrograms per gram; This indicates the first [number] in the cross-boundary concentration distribution sequence. Concentration values ​​at each node, in micrograms per gram; The node pair number is denoted as 1 to the total number of nodes minus 1. For the cross-boundary concentration distribution sequence at the epidermal-cortical boundary, node pair numbers from 1 to 10 are used. The concentration difference between 10 adjacent node pairs is calculated for each pair, resulting in a set of first-order difference values ​​corresponding to that boundary. This set contains 10 first-order difference values ​​numbered sequentially from 1 to 10. Each first-order difference value reflects the direction and magnitude of the change in lead mass fraction between its corresponding adjacent pairs. If the first-order difference values ​​exhibit a systematic sign characteristic, such as all being negative, it indicates a monotonically decreasing systematic trend in concentration along the normal direction. This is the basis for determining the direction of the concentration gradient and locating concentration abrupt changes. The above first-order difference calculation is performed on the cross-boundary concentration distribution sequences at the cortex-phloem boundary, phloem-xylem boundary, and xylem-pithelial boundary, respectively, to obtain the set of first-order difference values ​​corresponding to each boundary.

[0042] Step 52: Based on the set of first-order difference values, calculate the geometric rate of change between adjacent first-order difference values ​​to obtain a sequence of difference value change rates. Specifically, this includes: calculating the geometric rate of change between adjacent first-order difference values ​​according to the set of first-order difference values ​​for each sequence. The geometric rate of change reflects the rate of change of the first-order difference value itself along the spatial direction, i.e., the second-order change characteristic of the concentration distribution curve, used to characterize the acceleration of concentration change. The formula for calculating the geometric rate of change between adjacent first-order difference values ​​for a given set of first-order difference values ​​is: ; In the formula, Indicates the first The geometric rate of change of a pair of adjacent first-order difference values, in micrograms per gram per micrometer. Positive values ​​increase the first-order difference value along the positive direction of the normal, and negative values ​​decrease the first-order difference value along the positive direction of the normal. Indicates the first Each first-order difference value is expressed in micrograms per gram. Indicates the first Each first-order difference value is expressed in micrograms per gram. This represents the distance between adjacent spatial nodes, with a value of 15 micrometers, which is equal to half the spatial step size of the laser ablation scan. The value represents the rate of change index, ranging from 1 to the total number of nodes minus 2. For the set of first-order difference values ​​at the boundary between the epidermis and cortex, which contains 10 first-order difference values, the rate of change index is taken as 1 to 9. The geometric rate of change is calculated for each pair of 9 adjacent first-order difference value pairs, resulting in a sequence of difference value rate of change corresponding to that boundary. This sequence contains 9 geometric rate of change values ​​numbered sequentially from 1 to 9. Each element in this sequence quantitatively describes the second-order change of the first-order difference value associated with each spatial step of 15 micrometers along the cross-boundary direction. The node location with a larger absolute value of geometric rate of change indicates that the curvature of the concentration distribution curve reaches its maximum at that location, i.e., this is a potential anomaly location where the concentration change rate changes abruptly. The corresponding difference value rate of change sequences are calculated for the first-order difference value sets of the boundaries between cortex and phloem, phloem and xylem, and xylem and pith, respectively.

[0043] Step 53: In the differential rate of change sequence, identify nodes whose absolute value of the rate of change exceeds a preset threshold and locate the extreme point of concentration mutation. Specifically, this includes: based on each differential rate of change sequence, identifying nodes whose absolute value of the rate of change exceeds a preset threshold using a threshold determination method, and locating the extreme point of concentration mutation. The preset threshold is 0.100 micrograms per gram per micrometer. The physical meaning of this threshold is: within a spatial step of 15 micrometers, a change of at least 1.5 micrograms per gram in the first-order differential value is considered a concentration mutation with statistical and practical physical significance. The determination rule is: for each element in the differential rate of change sequence, calculate the absolute value of the element. When the absolute value is greater than or equal to the preset threshold, the spatial location corresponding to the element is marked as a candidate point for concentration mutation. When the absolute value is less than the preset threshold, the node does not meet the mutation determination condition and is not marked. For the sequence of the rate of change of the difference values ​​between the epidermis and the cortex boundary, the absolute value of each element is calculated and compared with a preset threshold. The spatial node positions corresponding to the elements that meet the judgment conditions are recorded as a set of concentration mutation candidate points. The spatial position corresponding to the mutation candidate point is the corresponding node in the cross-boundary concentration distribution sequence corresponding to the boundary.

[0044] For spatially adjacent mutation candidate points, where adjacent is defined as a distance of no more than one spatial step (15 micrometers), a further adjacent point merging process is performed: when the spatial distance between two adjacent mutation candidate points is no more than 15 micrometers, the node with the larger absolute value of the rate of change is retained as the final concentration mutation extreme point, while the node with the smaller absolute value of the rate of change is discarded. The physical basis for the merging process is that adjacent mutation candidate points are likely to correspond to the broadening effect of the same concentration mutation event on continuous spatial nodes. Merging them into a single point can avoid the repeated counting of the same mutation event. After threshold determination and adjacent point merging, a set of concentration mutation extreme points at the boundary between the epidermis and the cortex is obtained. Each concentration mutation extreme point is identified by its node index number in the cross-boundary concentration distribution sequence corresponding to the boundary and its corresponding actual physical coordinates. The same threshold determination and merging process is performed on the difference value change rate sequences of the cortex and phloem boundary, the phloem and xylem boundary, and the xylem and pith boundary, respectively, to obtain the set of concentration mutation extreme points corresponding to each boundary.

[0045] In this embodiment of the invention, a set of concentration difference values ​​is obtained by performing first-order difference calculation on adjacent spatial nodes in a cross-boundary concentration distribution sequence, and then a second-order geometric rate of change is calculated on the first-order difference value set to obtain a sequence of difference value change rates. By using a preset threshold to identify nodes in the difference value change rate sequence whose absolute value of change rate exceeds the limit, the extreme point of concentration mutation is located. This overcomes the technical defects of directly relying on the original concentration value, which makes it difficult to distinguish between the two concentration change modes of gradual diffusion and abrupt enrichment; relying solely on the first-order concentration difference value cannot eliminate the interference of background concentration fluctuations on the identification of mutation points; and manually interpreting the inflection point of the concentration distribution curve is highly subjective and cannot be quantified. This achieves the technical effect of automatically identifying and accurately locating the spatial location of extreme points of concentration mutation in a cross-tissue boundary concentration distribution sequence.

[0046] In a preferred embodiment of the present invention, step 6 above may include: Step 61: Extract concentration data within the preset neighborhood on both sides of the extreme concentration mutation point, and calculate the magnitude of the difference in concentration data between the two sides to obtain the concentration jump rate across the boundary of adjacent plant tissues. Specifically, this includes: calculating the magnitude of the difference in concentration within the preset neighborhood on both sides of each extreme concentration mutation point to obtain the concentration jump rate across the boundary of adjacent plant tissues. The preset neighborhood width is 5 spatial nodes. Taking a certain extreme concentration mutation point as an example, this extreme point is located at a node index position in a certain cross-boundary concentration distribution sequence. Extract the concentration data within the neighborhood on both sides of this extreme point: the left neighborhood takes the preset neighborhood width of nodes immediately adjacent to the extreme point along the negative normal direction. If the number of nodes on the left is less than the preset neighborhood width, then all existing nodes are taken; the right neighborhood takes the preset neighborhood width of nodes immediately adjacent to the extreme point along the positive normal direction. If the number of nodes on the right is less than the preset neighborhood width, then all existing nodes are taken. The formula for calculating the concentration jump rate is: ; In the formula, This represents the concentration jump rate at the extreme point of concentration abrupt change, expressed as a percentage. This represents the arithmetic mean of the concentration values ​​of all nodes in the right neighborhood, i.e., the side with the positive direction of the normal, in micrograms per gram. This represents the arithmetic mean of the concentration values ​​of all nodes in the left neighborhood, i.e., the side with the negative normal direction, in micrograms per gram. The value is a very small positive number, taken as 0.001 micrograms per gram, to avoid the denominator approaching zero and causing abnormal calculation results. For all extreme concentration abrupt changes at each boundary, the corresponding concentration jump rate is calculated by substituting into the above formula. The maximum concentration jump rate of each boundary is taken as the representative concentration jump rate of that boundary, and is denoted as: the maximum concentration jump rate at the boundary between the epidermis and cortex, the maximum concentration jump rate at the boundary between the cortex and phloem, the maximum concentration jump rate at the boundary between the phloem and xylem, and the maximum concentration jump rate at the boundary between the xylem and pith. The concentration jump rate across boundaries is a key indicator for distinguishing between exogenous surface pollution and endogenous background enrichment: exogenous surface pollution usually produces a large concentration jump at the boundary between the epidermis and cortex, while the jump rate of endogenous background enrichment is relatively small at each boundary.

[0047] To verify the effectiveness of concentration jump rate as an indicator for identifying pollution sources, 20 Astragalus membranaceus (Huangqi) slices from known pollution sources were selected for verification experiments. These included 10 samples from exogenous surface contamination and 10 samples from endogenous background enrichment. The exogenous contamination samples were prepared by soaking Astragalus membranaceus slices in a lead standard solution of known concentration, while the endogenous enrichment samples were taken from Astragalus membranaceus grown in soil known to be contaminated with lead. The detection method of this invention was applied to all samples, and the maximum concentration jump rate at the boundary between the epidermis and cortex was calculated. The values ​​for this indicator in the exogenous contamination samples were all greater than 85%, while the values ​​for the endogenous enrichment samples were all less than 25%. There was no overlap between the two sets of data, demonstrating that this indicator has the ability to distinguish pollution sources.

[0048] Step 62: Based on the concentration gradient feature values ​​and concentration jump rates, construct a heavy metal micro-region distribution feature vector. Specifically, this includes: constructing a heavy metal micro-region distribution feature vector based on five concentration gradient feature values, namely, epidermal concentration gradient feature values, cortex concentration gradient feature values, phloem concentration gradient feature values, xylem concentration gradient feature values, and pith concentration gradient feature values, as well as representative concentration jump rates of four boundaries, namely, the maximum concentration jump rate at the boundary between epidermis and cortex, the maximum concentration jump rate at the boundary between cortex and phloem, the maximum concentration jump rate at the boundary between phloem and xylem, and the maximum concentration jump rate at the boundary between xylem and pith, and combining these with the overall concentration levels of each tissue region. This feature vector contains 11 feature components, characterizing the micro-spatial distribution of lead in the cross-section of Astragalus membranaceus slices from three dimensions: concentration gradient within the tissue, intensity of concentration jumps across tissue boundaries, and overall concentration level in the tissue region. The components of the feature vector are defined as follows: Components 1 to 5: concentration gradient feature values ​​for the five tissue regions (epidermis, cortex, phloem, xylem, and pith), in micrograms per gram per micrometer; Components 6 to 9: maximum concentration jump rates at the boundaries of epidermis and cortex, cortex and phloem, phloem and xylem, and xylem and pith, in percentage; Component 10: ... The average concentration in the cortex region is defined as the arithmetic mean of the lead mass fraction of all spatial nodes in the region corresponding to label value 1, in micrograms per gram. The eleventh component is the average concentration in the xylem region, defined as the arithmetic mean of the lead mass fraction of all spatial nodes in the region corresponding to label value 4, in micrograms per gram. All 11 feature components are objective quantitative indicators calculated from experimental data, without relying on subjective judgment or experience. This feature vector compresses the high-dimensional two-dimensional spatial concentration distribution data into an 11-dimensional low-dimensional representation vector. The two-dimensional spatial concentration distribution data contains the concentration values ​​of thousands of spatial nodes and pixel-level annotation information of micro-region tissue morphology masks.

[0049] Step 63 involves matching the heavy metal micro-region distribution feature vector with preset heavy metal pollution source tracing rules to determine the pollution source type of the target heavy metal element, obtaining detection results characterizing the spatial distribution state and pollution source type of the heavy metal micro-region. Specifically, this includes matching each of the 11 feature components of the heavy metal micro-region distribution feature vector with the conditions of the preset heavy metal pollution source tracing rules to determine the pollution source type of the target heavy metal element. The preset heavy metal pollution source tracing rules contain three mutually exclusive and complete rules, corresponding to three different pollution source types. All thresholds involved in the above rules are predetermined using a known sample calibration method. The specific calibration process is as follows: For the first gradient threshold, first jump rate threshold, and first ratio threshold corresponding to the determination of exogenous surface contamination, 10 Astragalus membranaceus slices samples confirmed by independent traceability methods to be exogenous surface contamination were selected. The sample preparation, imaging, scanning, and feature calculation process was fully executed according to this method to obtain the feature component values ​​corresponding to each sample. The minimum value of the epidermal layer concentration gradient feature value, the minimum value of the maximum value of the concentration jump rate at the epidermal and cortex boundary, and the minimum value of the ratio of the average epidermal layer concentration to the average xylem concentration in the 10 samples were respectively taken as the corresponding three thresholds to ensure the sensitivity of exogenous surface contamination identification. In this embodiment, the calibrated first gradient threshold value was 0.050 micrograms per gram per micrometer, the first jump rate threshold value was 80.0%, and the first ratio threshold value was 3.0.

[0050] For the second gradient threshold, second jump rate threshold, and second ratio threshold corresponding to the determination of endogenous background enrichment, 10 Astragalus membranaceus slices samples confirmed by independent traceability methods to be enriched in endogenous background were selected. The sample preparation, imaging, scanning, and feature calculation processes were fully executed according to this method to obtain the corresponding feature component values ​​for each sample. The maximum value of the epidermal concentration gradient feature value, the maximum value of the epidermal-cortex boundary concentration jump rate, and the maximum value of the ratio of the mean epidermal concentration to the mean xylem concentration in the 10 samples were taken as the corresponding three thresholds to ensure the specificity of endogenous background enrichment identification. The second gradient threshold was set to 0.020 μg / g / μm, the second jump rate threshold to 30.0%, and the second ratio threshold to 1.5. The calibration thresholds are applicable to the detection of lead and Astragalus membranaceus slices. For other heavy metal elements or other plant processed products, samples of the corresponding known pollution types should be selected according to the same calibration method, and the corresponding thresholds should be re-determined. The judgment logic of the rule for determining exogenous surface pollution is a logical AND operation of three conditions. When conditions a, b, and c are satisfied at the same time, the pollution source type is determined to be exogenous surface pollution.

[0051] Condition a: The first component of this feature vector, namely the epidermal concentration gradient feature value, is greater than or equal to the first gradient threshold, which is 0.050 micrograms per gram per micrometer. Physical criterion: Exogenous pollutants mainly enter plant tissues through surface adhesion or infiltration diffusion from the surface inwards. As the outermost tissue, the epidermis is the first to come into contact with and trap exogenous heavy metals, forming a steep concentration decreasing gradient from the outside to the inside. Therefore, the concentration gradient value inside the epidermis is relatively high. Condition b: The sixth component of this feature vector, namely the maximum concentration jump rate at the boundary between the epidermis and cortex, is greater than or equal to the first jump rate threshold, which is 80.0%. Physical criterion: The dense keratinized structure of the epidermis forms a physical barrier to the inward diffusion of heavy metals. The exogenous heavy metals enriched in the epidermis and the relatively low background concentration in the cortex form a concentration cliff at the boundary. Condition c: The ratio of the tenth component to the eleventh component of the feature vector, i.e. the ratio of the average concentration of the epidermis to the average concentration of the xylem, is greater than or equal to the first ratio threshold, which is 3.0. Physical criteria: Exogenous pollution is characterized by a distribution pattern of enrichment in the surface layer and lower concentration in the inner layer. The concentration difference between the epidermis and xylem is significant. The judgment logic of the endogenous background enrichment judgment rule is a logical AND operation of three conditions. When conditions d, e, and f are satisfied at the same time, the pollution source type is determined to be endogenous background enrichment.

[0052] Condition d: The first component of this eigenvector, i.e., the epidermal concentration gradient characteristic value, is less than or equal to the second gradient threshold, which is 0.020 micrograms per gram per micrometer. Physical criterion: After being absorbed from the soil by the roots, endogenously enriched heavy metals are transported upwards through the xylem vessels and distributed to various tissues via transpiration. As a non-metabolic active tissue, the epidermis does not have a preferential enrichment mechanism, therefore the concentration gradient within the epidermis is gentle. Condition e: The sixth component of this eigenvector, i.e., the maximum concentration jump rate at the boundary between the epidermis and cortex, is less than or equal to the second jump rate threshold, which is 30.0%. Physical criterion: After long-distance transport and transmembrane translocation through the vascular system, endogenous heavy metals tend to be evenly distributed among various tissues within the plant, with insignificant cross-boundary jumps. Condition f: The ratio of the tenth component to the eleventh component of the feature vector, i.e. the ratio of the epidermal to the xylem concentration, is less than or equal to the second ratio threshold, which is 1.5. Physical criterion: The concentration difference among tissues is small under endogenous enrichment.

[0053] The determination rule for tissue-specific accumulation is as follows: when the feature vector does not satisfy all three conditions of rule 1, and does not simultaneously satisfy all three conditions of rule 2, the source of contamination is determined to be tissue-specific accumulation. This rule corresponds to the selective binding or compartmentalized enrichment mechanism of certain heavy metal ions in specific plant tissues, leading to local anomalies in concentration distribution. However, this anomaly pattern does not conform to typical exogenous input characteristics, nor does it conform to uniform endogenous distribution characteristics. Specific tissues include tissue regions with special chemical composition and structure, such as the phloem sieve tube companion cell complex or the cell walls of xylem vessels. The 11 component values ​​of the feature vector are matched sequentially with the conditions of rules 1 to 3. Through logical judgment, a conclusion is given on the pollution source type of the target heavy metal element. The conclusion is one of three: exogenous surface pollution, endogenous background enrichment, or tissue-specific accumulation. The final output detection result includes two components: the first part is the two-dimensional spatial concentration distribution matrix of the target heavy metal element on the micro-area slice to be tested and the micro-area tissue morphology mask, used to characterize the spatial distribution state of the micro-area; the second part is the pollution source type determined after matching with the source tracing judgment rule, used to characterize the source tracing qualitative conclusion of the heavy metal source. The two parts of information together constitute a complete detection report on the micro-area distribution state of heavy metals and its pollution causes in root and rhizome Chinese medicinal slices.

[0054] In this embodiment of the invention, the concentration data within a preset neighborhood on both sides of the extreme concentration mutation point is extracted to calculate the concentration jump rate across the boundary of adjacent plant tissues. The concentration gradient feature value and the concentration jump rate are then fused to construct a heavy metal micro-region distribution feature vector. This feature vector is then matched with preset heavy metal pollution source tracing rules to determine the pollution source type. Therefore, this method overcomes the technical defects of conventional heavy metal total detection methods, which can only obtain the overall average concentration and cannot trace the pollution source, lack spatial distribution dimension information of a single concentration indicator leading to insufficient basis for source determination, and rely on experience interpretation without objective quantitative judgment standards. This method achieves the technical effect of extracting multidimensional quantitative indicators from the spatial distribution characteristics of micro-regions and fusing them into a feature vector, realizing the automated and objective determination of exogenous / endogenous pollution types through rule matching, and finally outputting a complete detection result that contains both information on the spatial distribution status of heavy metal micro-regions and the pollution source type.

[0055] like Figure 3 As shown, embodiments of the present invention also provide a system for detecting the micro-regional distribution of heavy metals in processed plant products, comprising: The acquisition module is used to acquire cross-sectional micro-area slice samples of root and rhizome Chinese medicinal materials, and to perform cryo-embedding and ultrathin sectioning and fixation on the surface of the cross-sectional micro-area slice samples to obtain the micro-area slice to be tested. The extraction module is used to perform bright-dark field composite optical microscopy imaging on the micro-area slice to be tested, to obtain an initial optical image containing plant tissue morphology features, and to extract features from the initial optical image to obtain a micro-area tissue morphology mask. The analysis module is used to perform laser ablation micro-area elemental surface scanning analysis on the micro-area slice to be tested based on the target scanning area defined by the micro-area tissue morphology mask, and obtain the two-dimensional spatial concentration distribution matrix of the target heavy metal elements on the micro-area slice to be tested; The module is used to perform spatial coordinate registration and overlay mapping between the two-dimensional spatial concentration distribution matrix and the micro-region tissue morphology mask to obtain the concentration gradient feature values ​​of each plant tissue region; based on the micro-region tissue morphology mask, the geometric contour of the boundary of adjacent plant tissues is determined, and the corresponding concentration data in the two-dimensional spatial concentration distribution matrix is ​​extracted along the normal direction of the geometric contour to construct the cross-boundary concentration distribution sequence. The calculation module is used to calculate the first-order difference value of adjacent spatial nodes in the concentration distribution sequence; and to locate the extreme points of concentration abrupt changes by calculating the geometric rate of change of the first-order difference value. The matching module is used to obtain the concentration jump rate across the boundary of adjacent plant tissues by calculating the change range of the concentration difference in the preset neighborhood on both sides of the extreme value of concentration jump; based on the concentration gradient feature value and the concentration jump rate, it matches the preset heavy metal pollution source tracing judgment rules to obtain the detection results.

[0056] It should be noted that this system is a system corresponding to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0057] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0058] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0059] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the 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 detecting the micro-regional distribution of heavy metals in processed plant products, characterized in that, The method includes: Cross-sectional micro-area slices of rhizome-type Chinese medicinal materials were obtained, and the surface of the cross-sectional micro-area slices was subjected to cryo-embedding and ultrathin sectioning and fixation to obtain the micro-area slices to be tested. Bright-dark field composite optical microscopy imaging was performed on the section of the micro-area to be measured to obtain an initial optical image containing plant tissue morphology features. Feature extraction was performed on the initial optical image to obtain a micro-area tissue morphology mask. Based on the target scanning area defined by the micro-region morphology mask, laser ablation micro-region elemental surface scanning analysis is performed on the micro-region slice to be tested to obtain the two-dimensional spatial concentration distribution matrix of the target heavy metal elements on the micro-region slice to be tested. The concentration gradient feature values ​​of each plant tissue region are obtained by spatial coordinate registration and overlay mapping of the two-dimensional spatial concentration distribution matrix and the micro-region tissue morphology mask. Based on the micro-region tissue morphology mask, the geometric contour of the boundary of adjacent plant tissues is determined, and the corresponding concentration data in the two-dimensional spatial concentration distribution matrix is ​​extracted along the normal direction of the geometric contour to construct the cross-boundary concentration distribution sequence. Calculate the first-order difference between adjacent spatial nodes in the concentration distribution sequence; locate the extreme points of concentration abrupt changes by calculating the geometric rate of change of the first-order difference. By calculating the magnitude of the concentration difference within a preset neighborhood on both sides of the extreme concentration mutation point, the concentration jump rate across the boundary of adjacent plant tissues is obtained; based on the concentration gradient characteristic value and the concentration jump rate, the preset heavy metal pollution source tracing judgment rules are matched to obtain the detection results.

2. The method for detecting the micro-regional distribution of heavy metals in processed plant products according to claim 1, characterized in that, Cross-sectional micro-section samples of rhizome-type Chinese medicinal herbs were obtained, and the surfaces of the cross-sectional micro-section samples were subjected to cryo-embedding and ultrathin sectioning fixation to obtain the micro-sections to be tested, including: Rhizome-type Chinese medicinal materials were processed into slices, and these slices were subjected to liquid nitrogen flash freezing to obtain cryo-solidified samples. The cryo-solidified samples were then cryo-embedded in low-temperature resin to obtain tissue embedding blocks. The tissue embedding block was transversely cut using an ultramicrotome to obtain cross-sectional microsection samples. The surface of the cross-sectional micro-area slice sample is smoothed and its morphology is fixed to eliminate the micro-morphological undulations on the slice surface, thus obtaining the micro-area slice to be tested.

3. The method for detecting the micro-regional distribution of heavy metals in processed plant products according to claim 2, characterized in that, Bright-field and dark-field composite optical microscopy imaging was performed on the section of the micro-region to be measured to obtain an initial optical image containing plant tissue morphology features. Feature extraction was performed on the initial optical image to obtain a micro-region tissue morphology mask, including: Multi-channel imaging was performed on the micro-area section to be tested using a bright-dark field composite optical microscope to obtain an initial optical image containing plant tissue morphological features; the initial optical image was then processed to grayscale and contrast enhancement to obtain an enhanced grayscale image. Edge detection and morphological closing operations are performed on the enhanced grayscale image to extract the regional boundary contours of the epidermis, cortex, phloem and xylem, thus obtaining the initial boundary contour map; The initial boundary contour map is subjected to connected domain geometric segmentation and region filling to obtain a micro-region tissue morphology mask containing spatial location information of each tissue region.

4. The method for detecting the micro-regional distribution of heavy metals in processed plant products according to claim 3, characterized in that, Based on the target scanning area defined by the micro-region morphology mask, laser ablation micro-area elemental surface scanning analysis is performed on the micro-area slice to be tested, obtaining the two-dimensional spatial concentration distribution matrix of the target heavy metal elements on the micro-area slice, including: Extract the spatial coordinate range of each plant tissue region in the micro-area tissue morphology mask to obtain the target scanning area; Based on the target scanning area, the laser ablation device is controlled to perform point-by-point array ablation on the micro-area slice to be tested, so as to obtain an aerosol sample stream. The aerosol sample stream was analyzed by mass spectrometry to collect elemental isotope signals, and the signal intensity values ​​of the target heavy metal elements at each erosion site were obtained. The signal intensity values ​​of the target heavy metal elements at each erosion point are quantitatively converted by combining them with a preset standard material calibration curve to obtain a two-dimensional spatial concentration distribution matrix of the target heavy metal elements on the micro-area slice to be tested.

5. The method for detecting the micro-regional distribution of heavy metals in processed plant products according to claim 4, characterized in that, By performing spatial coordinate registration and overlay mapping between the two-dimensional spatial concentration distribution matrix and the micro-region tissue morphology mask, the concentration gradient characteristic values ​​of each plant tissue region are obtained. Based on micro-region tissue morphology masks, the geometric contours of adjacent plant tissue boundaries are determined, and corresponding concentration data in the two-dimensional spatial concentration distribution matrix are extracted along the normal direction of the geometric contours to construct a cross-boundary concentration distribution sequence, including: Spatial coordinate registration and overlay mapping are performed on the two-dimensional spatial concentration distribution matrix and the micro-region tissue morphology mask to obtain a spatially aligned composite mapping matrix; Based on the composite mapping matrix, the concentration gradient feature values ​​of the target heavy metal element in each plant tissue region are calculated; the geometric contours of the boundaries between adjacent plant tissues in the micro-region tissue morphology mask are extracted to obtain the boundary contour lines. Along the normal direction of the boundary contour line, extract the corresponding concentration data from the two-dimensional spatial concentration distribution matrix to construct a cross-boundary concentration distribution sequence.

6. The method for detecting the micro-regional distribution of heavy metals in processed plant products according to claim 5, characterized in that, Calculate the first-order difference between adjacent spatial nodes in the concentration distribution sequence; The extreme points of concentration abrupt changes are located by calculating the geometric rate of change of the first-order difference, including: Calculate the concentration difference between adjacent spatial nodes in the cross-boundary concentration distribution sequence to obtain the set of first-order difference values; Based on the set of first-order difference values, the geometric rate of change between adjacent first-order difference values ​​is calculated to obtain the sequence of difference value change rates. In the differential rate of change sequence, nodes whose absolute value of the rate of change exceeds a preset threshold are identified, and extreme points of concentration mutation are located.

7. The method for detecting the micro-regional distribution of heavy metals in processed plant products according to claim 6, characterized in that, By calculating the concentration difference within a preset neighborhood on either side of the extreme concentration abrupt change point, the concentration jump rate across adjacent plant tissue boundaries is obtained. Based on the concentration gradient characteristic value and the concentration jump rate, a preset heavy metal pollution source tracing judgment rule is matched to obtain the detection results, including: Extract concentration data within a preset neighborhood on both sides of the extreme concentration mutation point, and calculate the magnitude of the difference in concentration data on both sides to obtain the concentration jump rate across the boundary of adjacent plant tissues. Based on the concentration gradient eigenvalues ​​and concentration jump rate, a characteristic vector of heavy metal micro-region distribution is constructed. By matching the feature vector of heavy metal micro-region distribution with the preset heavy metal pollution source determination rules, the pollution source type of the target heavy metal element is determined, and the detection results characterizing the spatial distribution state and pollution source type of heavy metal micro-region are obtained.

8. A system for detecting the micro-regional distribution of heavy metals in processed plant products, wherein the system implements the method as described in any one of claims 1 to 7, characterized in that, include: The acquisition module is used to acquire cross-sectional micro-area slice samples of root and rhizome Chinese medicinal materials, and to perform cryo-embedding and ultrathin sectioning and fixation on the surface of the cross-sectional micro-area slice samples to obtain the micro-area slice to be tested. The extraction module is used to perform bright-dark field composite optical microscopy imaging on the micro-area slice to be tested, to obtain an initial optical image containing plant tissue morphology features, and to extract features from the initial optical image to obtain a micro-area tissue morphology mask. The analysis module is used to perform laser ablation micro-area elemental surface scanning analysis on the micro-area slice to be tested based on the target scanning area defined by the micro-area tissue morphology mask, and obtain the two-dimensional spatial concentration distribution matrix of the target heavy metal elements on the micro-area slice to be tested; A module is constructed to perform spatial coordinate registration and overlay mapping between the two-dimensional spatial concentration distribution matrix and the micro-region tissue morphology mask to obtain the concentration gradient feature values ​​of each plant tissue region. Based on the micro-region tissue morphology mask, the geometric contour of the boundary between adjacent plant tissues is determined, and the corresponding concentration data in the two-dimensional spatial concentration distribution matrix is ​​extracted along the normal direction of the geometric contour to construct a cross-boundary concentration distribution sequence. The calculation module is used to calculate the first-order difference value of adjacent spatial nodes in the concentration distribution sequence; and to locate the extreme points of concentration abrupt changes by calculating the geometric rate of change of the first-order difference value. The matching module is used to obtain the concentration jump rate across the boundary of adjacent plant tissues by calculating the change range of the concentration difference in the preset neighborhood on both sides of the extreme value of concentration jump; based on the concentration gradient feature value and the concentration jump rate, it matches the preset heavy metal pollution source tracing judgment rules to obtain the detection results.

9. A computing device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program that, when executed by a processor, implements the method as described in any one of claims 1 to 7.

Images