Fruit tracing method and system based on low-field magnetic nuclear magnetic resonance

Abstract

The invention relates to the technical field of traceability identification, in particular to a fruit traceability method and system based on low-field magnetic nuclear magnetic resonance, and the method comprises the steps: recognizing the color classification and appearance features of a fruit according to an image of the fruit; performing low-field magnetic nuclear magnetic resonance on the pulp separated from the fruits, and determining the sweetness and acidity of the pulp according to low-field magnetic resonance signals; obtaining a sweetness proportion range and an acidity proportion range according to color classification and the appearance characteristics; judging whether the sweetness is within a sweetness proportion range and whether the acidity is within an acidity proportion range, if so, judging that the appearance of the fruit is forged, and if not, determining the appearance producing area of the fruit; obtaining element contents of kernels separated from the fruits, and determining element producing areas of the fruits according to the element contents; and determining whether the fruits are counterfeited from the producing area according to whether the appearance producing area corresponds to the element producing area. According to the invention, accurate detection of appearance and production place counterfeiting by using a small amount of fruit samples is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of traceability and identification, and in particular to a fruit traceability method and system based on low-field nuclear magnetic resonance. Background Art

[0002] Low-field nuclear magnetic resonance technology (LF-NMR) generates nuclear magnetic resonance signals by allowing the low-energy nuclear magnetic moment to absorb the energy provided by the alternating field strength and then transition to a high-energy state under the interaction of constant field strength and alternating field strength. In the process of fruit testing, the nuclear magnetic resonance signal can reflect the distribution and state of hydrogen atoms inside the fruit sample. The water content, sugar and other components in the fruit all contain hydrogen atoms, and the water content and sugar content of the fruit are related to whether the fruit has browned, rotted and been damaged. Therefore, low-field nuclear magnetic resonance technology can be used to test the quality of fruit.

[0003] In the sales process, due to import tariffs and transportation costs, although imported fruits are similar in appearance to domestic fruits, their prices are quite different. Therefore, the traceability of the origin of fruits is of great significance to regulating the market. In addition, EU laws and regulations and my country's Food Safety Law have relevant provisions on the origin and traceability of fruits, requiring the establishment of a traceability system for fruit safety.

[0004] At present, the main methods of falsifying the origin of fruits are: adding dyes, preservatives, and sweeteners to make the flesh of low-priced domestic fruits close to the flesh of high-priced imported fruits, making it difficult to simply distinguish the fake fruits from their appearance and taste; falsifying labels, falsely marking the production date, shelf life, origin and other information of the fruit to deceive consumers.

[0005] Therefore, how to use fewer, high-priced imported fruit samples to conduct comprehensive traceability judgments on the above-mentioned appearance counterfeiting and origin counterfeiting, and to achieve accurate judgments on different fruit counterfeiting methods, is a technical problem that needs to be solved. Summary of the invention

[0006] To this end, the present invention provides a fruit traceability method and system based on low-field magnetic nuclear magnetic resonance. Through the image of the fruit and the low-field magnetic resonance signal of the pulp, it is possible to quickly determine the appearance of the fruit that is forged by adding dyes, preservatives, and sweeteners with inaccurate proportions. Through the comprehensive determination of the element content and image of the fruit core and the low-field magnetic resonance signal, it is possible to accurately detect origin fraud using a small amount of fruit samples.

[0007] To achieve the above object, the present invention proposes a fruit tracing method based on low-field magnetic nuclear magnetic resonance, comprising:

[0008] Acquire an image of the fruit, and identify the color classification and appearance characteristics of the fruit according to the image;

[0009] Performing low-field nuclear magnetic resonance on the pulp separated from the fruit to obtain a low-field magnetic resonance signal, and determining the sweetness and acidity of the pulp according to the low-field magnetic resonance signal;

[0010] Determining a sweetness ratio range and a sourness ratio range according to the color classification and the appearance characteristics;

[0011] Determine whether the sweetness is within the sweetness ratio range and whether the acidity is within the acidity ratio range, if so, determine that the appearance of the fruit is forged, if not, determine the appearance origin of the fruit;

[0012] Obtaining the element content of the fruit core separated from the fruit, and determining the element origin of the fruit according to the element content;

[0013] Whether the origin of the fruit is fake is determined based on whether the appearance origin and the element origin correspond.

[0014] Furthermore, the step of obtaining the sweetness ratio range and the acidity ratio range comprises:

[0015] Determining the purity, maturity and freshness of the fruit according to the color classification and the appearance characteristics;

[0016] Determining an initial sweetness ratio range and an initial acidity ratio range according to the variety purity and the maturity;

[0017] The initial sweetness ratio range and the initial acidity ratio range are adjusted according to the freshness to obtain the sweetness ratio range and the acidity ratio range.

[0018] Furthermore, the step of obtaining the color classification and the appearance characteristics includes:

[0019] Determining the appearance features and the color classification of the fruit in the image through a neural network-based image recognition model;

[0020] Determine the sub-variety of the fruit based on color type, stem shape, and skin appearance;

[0021] Among them, the appearance characteristics include the stem shape, stem brightness, epidermal appearance and epidermal glossiness, and the color classification includes the color type, characteristic chromaticity region area and characteristic chromaticity value.

[0022] Furthermore, the process of determining the variety purity of the fruit according to the color classification and the appearance characteristics is as follows:

[0023] The purity of the variety is determined by an image similarity model based on the color type, the stem shape, the epidermal appearance and the historical picture library of the origin corresponding to the subdivided variety.

[0024] Further, the step of determining the initial sweetness ratio range and the initial acidity ratio range according to the variety purity and the maturity includes:

[0025] Determining the initial sweetness ratio range and the initial acidity ratio range by a genetic algorithm based on the variety purity and the maturity;

[0026] Among them, the maturity is the area ratio of the characteristic chromaticity area to the area standard value, and the variety purity is the chromaticity ratio of the characteristic chromaticity value to the chromaticity standard value.

[0027] In the above scheme, the determination of the proportion range for judging acidity and sweetness is achieved by taking into account the differences in fruit samples and typical conditions of sub-varieties, making it more accurate to judge whether the appearance of fruits with the addition of multiple dyes, preservatives and sweeteners is forged to falsify the origin.

[0028] Furthermore, the step of determining the freshness of the fruit according to the color classification and the appearance characteristics comprises:

[0029] The stem brightness corresponds to a stem brightness level, the skin gloss corresponds to a gloss level, and the freshness is determined based on the stem brightness level and the gloss level.

[0030] Further, the process of determining the sweetness and acidity of the pulp according to the low-field magnetic resonance signal is as follows:

[0031] Determining the sweetness corresponding to the low-field magnetic resonance signal by characterizing a correlation model between the low-field magnetic resonance signal per unit mass and the sweetness of the fruit;

[0032] The acidity corresponding to the low-field magnetic resonance signal is determined by a fitting model that characterizes the transverse relaxation time and the acidity of the fruit.

[0033] In the above scheme, high-resolution signals using low-field magnetic nuclear magnetic resonance are realized, so that even if a small amount of colorants, preservatives or sweeteners are added to the fruit, it can be accurately detected.

[0034] Further, the elements used to determine the elemental origin of the fruit include potassium, cobalt, rubidium, nickel, manganese, calcium, hydrogen stable isotopes and oxygen stable isotopes;

[0035] Determining the origin of the element by cluster analysis of the element content;

[0036] The appearance origin is the continental origin, and the element origin is the national origin;

[0037] Determine whether the geographical scope of the appearance origin includes the geographical scope of the element origin. If so, the fruit does not have counterfeit origin. Otherwise, the fruit has counterfeit origin.

[0038] The present invention also provides a fruit tracing system based on low-field magnetic nuclear magnetic resonance, wherein the fruit tracing system is equipped with the fruit tracing method based on low-field magnetic nuclear magnetic resonance, and the fruit tracing system comprises:

[0039] Vision sensor to obtain images of the fruit;

[0040] A low-field magnetic resonance analyzer is used to perform low-field magnetic resonance on the fruit core to obtain low-field magnetic resonance signals;

[0041] Inductively coupled plasma mass spectrometer to obtain the elemental content of fruit cores;

[0042] A control device communicates with the visual sensor, the inductively coupled plasma mass spectrometer and the low-field magnetic resonance analyzer, and is used to determine whether the appearance and appearance origin of the fruit are forged based on the image and the low-field magnetic resonance signal, and determine whether the origin of the fruit is fake based on the appearance origin and the element content.

[0043] Furthermore, the control device is equipped with a detection system cloud platform to generate a fruit detection report including the image, the low-field magnetic resonance signal, the element content, whether the appearance is forged, and whether the origin is fake.

[0044] In the above solution, the centralized management of testing equipment through the cloud platform makes the testing process faster and more standardized.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. Through the image of the fruit and the low-field magnetic resonance signal of the pulp, it is possible to quickly determine the appearance of the fruit forged by adding dyes, preservatives, and sweeteners with inaccurate proportions. Through the comprehensive judgment of the element content, image, and low-field magnetic resonance signal of the fruit core, it is possible to accurately detect the origin fraud using a small amount of fruit samples.

[0047] 2. It has achieved the determination of the proportion range for judging acidity and sweetness, which takes into account the differences in fruit samples and typical conditions of sub-varieties, making it more accurate to judge the appearance of fruits that have been falsified to fake their origin by adding multiple coloring agents, preservatives, and sweeteners.

[0048] 3. The high-resolution signal of low-field magnetic nuclear magnetic resonance is realized, so that even if very small amounts of dyes, preservatives or sweeteners are added to the fruit, they can be accurately and quickly detected.

[0049] 4. The centralized management of testing equipment through the cloud platform makes the testing process faster and more standardized. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1 It is a schematic diagram of the general process of the fruit tracing method based on low-field magnetic nuclear magnetic resonance according to an embodiment of the present invention;

[0051] Figure 2 A schematic diagram of a process for determining counterfeit origin according to an embodiment of the present invention;

[0052] Figure 3 This is a schematic diagram of a correlation model fitting between a low-field magnetic resonance signal representing a unit mass and the sweetness of a fruit, taking cherries as an example in an embodiment of the present invention;

[0053] Figure 4 The figure is a schematic diagram of the general structure of a fruit traceability system based on low-field magnetic nuclear magnetic resonance according to an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0055] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0056] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0057] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0058] like Figures 1 to 4As shown, the present invention provides a fruit traceability method and system based on low-field magnetic nuclear magnetic resonance. Through the image of the fruit and the low-field magnetic resonance signal of the pulp, it is possible to quickly determine the appearance of the fruit forged by adding dyes, preservatives and sweeteners with inaccurate proportions. Through the comprehensive determination of the element content and image of the fruit core and the low-field magnetic resonance signal, it is possible to accurately detect the origin fraud using a small amount of fruit samples.

[0059] in, Figure 1 It is a schematic diagram of the general process of the fruit tracing method based on low-field magnetic nuclear magnetic resonance according to an embodiment of the present invention; Figure 2 A schematic diagram of a process for determining counterfeit origin according to an embodiment of the present invention; Figure 3 It is a schematic diagram of the correlation model fitting between the low-field magnetic resonance signal per unit mass and the sweetness of fruit according to an embodiment of the present invention; Figure 4 The figure is a schematic diagram of the general structure of a fruit traceability system based on low-field magnetic nuclear magnetic resonance according to an embodiment of the present invention.

[0060] like Figures 1 to 4 As shown, this embodiment proposes a fruit traceability method and system based on low-field magnetic nuclear magnetic resonance, the method comprising: acquiring an image of the fruit, identifying the color classification and appearance characteristics of the fruit according to the image; performing low-field magnetic nuclear magnetic resonance on the pulp separated from the fruit to obtain a low-field magnetic resonance signal, and determining the sweetness and acidity of the pulp according to the low-field magnetic resonance signal; obtaining a sweetness ratio range and an acidity ratio range according to the color classification and the appearance characteristics; judging whether the sweetness is within the sweetness ratio range and whether the acidity is within the acidity ratio range, if so, judging that the appearance of the fruit is forged, if not, determining the appearance origin of the fruit; acquiring the element content of the fruit core separated from the fruit, and determining the element origin of the fruit according to the element content; determining whether the origin of the fruit is forged according to whether the appearance origin corresponds to the element origin.

[0061] It should be noted that the fruits described in this embodiment are all fruits with pits, and their scientific name is drupe.

[0062] It should be noted that when it is determined that the appearance of the fruit is forged, the testing can be stopped and the fruit quality / origin traceability can be judged as unqualified. There is no need to conduct more complex element content testing on the fruit, which speeds up the testing efficiency.

[0063] Furthermore, the step of obtaining the sweetness ratio range and the acidity ratio range includes: determining the variety purity, maturity and freshness of the fruit according to the color classification and the appearance characteristics; determining an initial sweetness ratio range and an initial acidity ratio range according to the variety purity and the maturity; adjusting the initial sweetness ratio range and the initial acidity ratio range according to the freshness to obtain the sweetness ratio range and the acidity ratio range.

[0064] Furthermore, the step of obtaining the color classification and the appearance characteristics includes: determining the appearance characteristics and the color classification of the fruit in the image through an image recognition model based on a neural network; determining the variety of the fruit according to the color type, stem shape and skin shape; wherein the appearance characteristics include the stem shape, stem brightness, skin shape and skin gloss, and the color classification includes the color type, characteristic chromaticity region area and characteristic chromaticity value.

[0065] Specifically, the image recognition model is realized by the convolutional neural network based on YOLOv8 and Kitti data set. YOLOv8 is the next generation algorithm model developed by Ultra Lyticics after the YOLOv5 algorithm, and currently supports image classification, object detection and instance segmentation tasks. The image recognition model includes: a new backbone network, a new Ancher-Free detection head and a new loss function. Therefore, the present embodiment utilizes the YOLOv8 target detection algorithm to realize a fruit database system based on the YOLOV8 model and Kitti data set, and then uses the Pys ide6 library to build an interface system to complete the development of the fruit detection page. By adjusting the detection confidence threshold and the IOU threshold, the accuracy of the detection is more suitable.

[0066] The Kitt i data set used in this embodiment marks the six categories of fruit stem morphology, stem brightness, epidermal shape, epidermal glossiness, color type, and fruit variety, and the data set contains 9450 pictures in total. The categories in this data set have a large number of rotations and different lighting conditions, which helps to train a more robust detection model. The Kitt i detection and recognition data set of this paper experiment contains 8000 pictures in the training set and 1450 pictures in the verification set. Since the YOLOv8 algorithm has restrictions on the input picture size, it is necessary to adjust all pictures to the same size. In order to reduce the distortion of the picture as much as possible without affecting the detection accuracy, all pictures are adjusted to a size of 640x640 and the original aspect ratio is maintained. In addition, in order to enhance the generalization ability and robustness of the model, data enhancement techniques are used, including random rotation, scaling, cropping and color conversion, etc., to expand the data set and reduce the risk of overfitting. After identifying the fruit variety, the image recognition model runs again to identify the characteristic chromaticity area area and the characteristic chromaticity value of the fruit.

[0067] Furthermore, the step of determining the variety purity of the fruit based on the color classification and the appearance characteristics includes: determining the variety purity by comparing the color type, the stem shape, the skin shape with the historical image library of the origin corresponding to the variety through an image similarity model.

[0068] Specifically, each fruit variety is provided with an image similarity model, and the image library of each image similarity model has a plurality of fruit images of different maturity and freshness. The image similarity model is implemented using the image encoder of the CLIP model: the image is input into the image encoder of the CLIP model to obtain an image embedding. The cosine similarity between the image embedding and all image embeddings of the origin history picture library is calculated, and the cosine similarity is used as the fruit similarity. The cosine similarity is the image vector of the image embedding multiplied by the image vector of the image embedding, divided by its vector modulus.

[0069] The variety purity refers to the genetic purity of the fruit variety, that is, the degree of consistency of its genetic characteristics with a specific variety. The higher the variety purity of the fruit, the closer the sweetness and acidity of the fruit are to purebred fruit. The resulting sweetness ratio range and acidity ratio range are close to the individual fruit.

[0070] Furthermore, the step of determining the initial sweetness ratio range and the initial acidity ratio range according to the variety purity and the maturity includes: determining the initial sweetness ratio range and the initial acidity ratio range according to the variety purity and the maturity through a genetic algorithm; wherein the maturity is the area ratio of the characteristic chromaticity region to the area standard value, and the variety purity is the chromaticity ratio of the characteristic chromaticity value to the chromaticity standard value.

[0071] It is understandable that as the fruit gradually matures, its color will gradually change from green to purple, red, yellow or white, which is characteristic of the fruit variety, and the color will gradually become uniform, the flesh will gradually become full and juicy, and the sweetness will gradually reach the highest value. Generally speaking, the darker the color of the fruit, the higher the sweetness (sugar content). Therefore, the ratio of the characteristic chromaticity area to the area standard value and the ratio of the characteristic chromaticity value to the chromaticity standard value can reflect the maturity of the fruit, and the higher the maturity, the higher the sweetness of the fruit.

[0072] According to the detection of multiple fruits of different varieties and purities during the growth process by handheld sugar content tester and handheld acidity tester, it was found that the maturity, variety purity, sweetness and acidity can be fitted to generate a linear correspondence curve, and there is a linear correspondence. Therefore, it is inferred that the sweetness ratio range and acidity ratio range of fruits of different maturity can be determined by the variety purity, area ratio and chromaticity ratio.

[0073] Specifically, the sweetness and acidity of fruits with variety purity during the growth process are recorded and input into a genetic algorithm (BP), and the optimal objective function is obtained through evolutionary calculation of the genetic algorithm. For example, by using multiple groups of sweetness, the fruit similarity, the area ratio of the characteristic chromaticity region to the area standard value, and the chromaticity ratio of the characteristic chromaticity value to the chromaticity standard value as training parameters, the optimal objective function can be obtained, for example: y1=arcctg(cos(x1 / (x1+x2+0.0001))+0.0023)×cos(x1+x2+x3+0.0015)), y2=arcctg(cos(x1 / (x1+x2+0.0025))+0.0023)×cos(x1+x2+x3+0.0035)), wherein y1 is the predicted sweetness standard value, y2 is the predicted acidity standard value, x1, x2 and x3 are the variety purity, the area ratio, and the chromaticity ratio, respectively. The predicted sweetness standard value is added or subtracted from the set value to serve as the initial sweetness ratio range of the sweetness, wherein the set value is preferably 5%, and may also be adjusted according to the typicality of the fruit variety, and may not exceed 10% at most. Similarly, the predicted acidity standard value is added or subtracted from the set value to serve as the initial acidity ratio range of the sweetness, wherein the set value is preferably 7%, and may also be adjusted according to the typicality of the fruit variety, and may not exceed 15% at most.

[0074] Specifically, the sub-categories of the fruit, and the corresponding color types, stem shapes, and skin shapes are: the color type of Red Fuji apple is red, the stem shape is woody, and the skin shape is smooth; the color type of Granny Smith apple is green, the stem shape is purple-brown, and the skin shape is oblate; the color type of Gala apple is yellow, the stem shape is smooth green leaves, and the skin shape is waxy.

[0075] The subdivided varieties can also be subdivided varieties of cherries, for example: Santina variety, the color type is dark red or purple-black, there is no stem morphology data, and the skin shape is heart-shaped and flat (because Santina's characteristics are obvious and clear, it can be distinguished only by the color type and skin shape); Regina variety, the color type is brown-red, date red or yellow-red color, there is no stem morphology data, and the skin shape is a larger circle; Brooks variety, the color type is light red, the stem morphology is relatively straight, and the skin shape is a smaller oval; Rainier variety, the color type is red and yellow or pink, there is no stem morphology data, and the skin shape is a larger ingot shape; Lapin variety, the color type is brown-red, and the skin shape is medium; Bing variety, the color type is dark red, the stem morphology is longer, and the skin shape is medium-round.

[0076] For example, the sweetness ratio of the Santina variety ranges from 14 to 22 in grams per 100 grams of pulp, and the acidity ratio of the Santina variety ranges from 7 to 13 in grams per 100 grams of pulp.

[0077] Furthermore, according to the color classification and the appearance characteristics, the step of determining the freshness of the fruit includes: corresponding the stem brightness to the stem brightness level, corresponding the skin glossiness to the glossiness level, and determining the freshness according to the stem brightness level and the glossiness level.

[0078] Specifically, the chromaticity ratio is the RGB ratio of the characteristic chromaticity value to the chromaticity standard value, the stem brightness is the RGB ratio of the RGB mean of the fruit stem color in the image to the RGB ratio of the standard tender green color, expressed in percentage. The epidermis glossiness is the reflectivity of the fruit surface, expressed in percentage.

[0079] Specifically, the stem brightness level includes stem level one to stem level five, which increase in proportion, and the proportions are 40% to 44%, 45% to 49% to more than 60%. The glossiness level includes glossiness level one to glossiness level five, which increase in proportion, and the proportions are 0% to 3%, 4% to 7% to more than 16%. When the stem brightness level is lower than stem level two, and the glossiness level is lower than glossiness level two, the freshness level is determined to be low; when the stem brightness level is greater than or equal to stem level three or the glossiness level is greater than or equal to glossiness level three, the freshness level is determined to be medium; when the stem brightness level is greater than or equal to stem level three and the glossiness level is greater than or equal to glossiness level three, the freshness level is determined to be high. For the low freshness level, the initial sweetness ratio range is reduced by 2 to obtain the sweetness ratio range, and the initial acidity ratio range is reduced by 0.5 to obtain the acidity ratio range. For the medium freshness level, the initial sweetness ratio range and the initial acidity ratio range remain unchanged to obtain the sweetness ratio range and the acidity ratio range. For a high freshness level, the sweetness ratio range is increased by 2, and the acidity ratio range is increased by 0.5, to obtain the sweetness ratio range and the acidity ratio range. It should be noted that the above process is preferably applied to the traceability detection of cherries.

[0080] Furthermore, the step of determining the sweetness and acidity of the fruit pulp according to the low-field magnetic resonance signal includes: determining the sweetness corresponding to the low-field magnetic resonance signal by characterizing a correlation model between the low-field magnetic resonance signal per unit mass and the sweetness of the fruit; and determining the acidity corresponding to the low-field magnetic resonance signal by characterizing a fitting model between the transverse relaxation time and the acidity of the fruit.

[0081] Specifically, a sugar content detector is used to detect the sweetness of the fruit, and it is determined that the amplitude of the low-field magnetic resonance signal (NMR) and the sweetness of the fruit have an inverse function relationship, and the inverse function is a correlation model, specifically y=-25x+88. In the formula, y is the sweetness, and x is the amplitude of the low-field magnetic resonance signal.

[0082] Specifically, an acidity detector is used to detect the acidity of the fruit, and it is determined that the amplitude of the transverse relaxation time (T2) has an inverse function relationship with the fruit acidity, and the inverse function is a correlation model, specifically y=-19x+21. In the formula, y is the acidity, and x is the transverse relaxation time.

[0083] In the above scheme, since the stem shape, skin appearance and color type of the fruit are strongly correlated with its variety and origin factors, the similarity with the fruit historical picture library is first determined, the proportion range is determined, and then the proportion range is adjusted according to the freshness of the fruit. In this way, accurate judgment of the counterfeit appearance of fruits with the addition of multiple coloring agents, preservatives and sweeteners is achieved.

[0084] Furthermore, the elements used to determine the element origin of the fruit include potassium, cobalt, rubidium, nickel, manganese, calcium, hydrogen stable isotopes and oxygen stable isotopes; the element origin is determined based on cluster analysis of the element content.

[0085] It is understandable that potassium, cobalt, rubidium, nickel, manganese, calcium, hydrogen stable isotopes and oxygen stable isotopes are related to the soil and climate of the fruit growing area. Therefore, through cluster analysis of them, the origin information of the fruit can be accurately determined.

[0086] Furthermore, the appearance origin is intercontinental origin information, and the element origin is national origin information; it is determined whether the geographical scope of the appearance origin includes the geographical scope of the element origin, if so, the fruit does not have origin fraud, otherwise, the fruit has origin fraud.

[0087] Specifically, the fruit variety corresponds to an intercontinental appearance origin, for example, the appearance origin of the Santina variety is South America, and the element origin is Chile.

[0088] The present embodiment also provides a fruit traceability system based on low-field magnetic nuclear magnetic resonance, the fruit traceability system is equipped with the fruit traceability method based on low-field magnetic nuclear magnetic resonance, the fruit traceability system comprises: a visual sensor for acquiring an image of the fruit; a low-field magnetic resonance analyzer for performing low-field magnetic resonance on the core of the fruit to acquire its low-field magnetic resonance signal; an inductively coupled plasma mass spectrometer for acquiring the element content of the core of the fruit; a control device for communicating with the visual sensor, the inductively coupled plasma mass spectrometer and the low-field magnetic resonance analyzer, for determining whether the appearance of the fruit and the appearance of the origin are forged according to the image and the low-field magnetic resonance signal, and determining whether the origin of the fruit is forged according to the appearance of the origin and the element content.

[0089] Furthermore, the control device is equipped with a detection system cloud platform to generate a fruit detection report including the image, the low-field magnetic resonance signal, the element content, whether the appearance is forged, and whether the origin is fake.

[0090] In the above solution, the centralized management of testing equipment through the cloud platform makes the testing process faster and more standardized.

[0091] Specifically, the control device is a server and a computer. The server controls the visual sensor, low-field magnetic resonance analyzer and inductively coupled plasma mass spectrometer through the Internet of Things, and is equipped with the various models described in this embodiment to achieve the judgment of traceability and fraud.

[0092] It can be understood that through the image of the fruit and the low-field magnetic resonance signal of the pulp, it is possible to quickly judge and detect the appearance of the fruit forged by adding dyes, preservatives, and sweeteners with inaccurate proportions. Through the comprehensive judgment of the element content, image, and low-field magnetic resonance signal of the fruit core, it is possible to accurately detect the origin fraud using a small amount of fruit samples. The determination of the ratio range for judging acidity and sweetness considering the differences in the fruit sample situation and the typical situation of the variety is achieved, making the judgment of the appearance of the fruit forged by adding multiple dyes, preservatives, and sweeteners to fake the origin more accurate. The high-resolution signal of low-field magnetic nuclear magnetic resonance is realized, so that even if a very small amount of dyes, preservatives or sweeteners is added to the fruit, it can be accurately and quickly detected. The centralized management of detection equipment through the cloud platform realizes the rapidity and standardization of the detection process.

[0093] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0094] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A fruit tracing method based on low-field magnetic nuclear magnetic resonance, characterized in that: include: Acquire an image of the fruit, and identify the color classification and appearance characteristics of the fruit according to the image; Performing low-field nuclear magnetic resonance on the pulp separated from the fruit to obtain a low-field magnetic resonance signal, and determining the sweetness and acidity of the pulp according to the low-field magnetic resonance signal; Determining a sweetness ratio range and a sourness ratio range according to the color classification and the appearance characteristics; Determine whether the sweetness is within the sweetness ratio range and whether the acidity is within the acidity ratio range, if so, determine that the appearance of the fruit is forged, if not, determine the appearance origin of the fruit; Obtaining the element content of the fruit core separated from the fruit, and determining the element origin of the fruit according to the element content; Whether the origin of the fruit is fake is determined based on whether the appearance origin and the element origin correspond.

2. The fruit tracing method based on low-field magnetic nuclear magnetic resonance according to claim 1, characterized in that: The steps of obtaining the sweetness ratio range and the acidity ratio range include: Determining the purity, maturity and freshness of the fruit according to the color classification and the appearance characteristics; Determining an initial sweetness ratio range and an initial acidity ratio range according to the variety purity and the maturity; The initial sweetness ratio range and the initial acidity ratio range are adjusted according to the freshness to obtain the sweetness ratio range and the acidity ratio range.

3. The fruit tracing method based on low-field magnetic nuclear magnetic resonance according to claim 2, characterized in that: The steps of obtaining the color classification and the appearance characteristics include: Determining the appearance features and the color classification of the fruit in the image through a neural network-based image recognition model; Determine the sub-variety of the fruit based on color type, stem shape, and skin appearance; Among them, the appearance characteristics include the stem shape, stem brightness, epidermal appearance and epidermal glossiness, and the color classification includes the color type, characteristic chromaticity region area and characteristic chromaticity value.

4. The fruit tracing method based on low-field magnetic nuclear magnetic resonance according to claim 3, characterized in that: The process of determining the purity of the fruit variety according to the color classification and the appearance characteristics is as follows: The purity of the variety is determined by an image similarity model based on the color type, the stem shape, the epidermal appearance and the historical picture library of the origin corresponding to the subdivided variety.

5. The fruit tracing method based on low-field magnetic nuclear magnetic resonance according to claim 3, characterized in that: The step of determining the initial sweetness ratio range and the initial acidity ratio range according to the variety purity and the maturity comprises: Determining the initial sweetness ratio range and the initial acidity ratio range by a genetic algorithm based on the variety purity and the maturity; Among them, the maturity is the area ratio of the characteristic chromaticity area to the area standard value, and the variety purity is the chromaticity ratio of the characteristic chromaticity value to the chromaticity standard value.

6. The fruit tracing method based on low-field magnetic nuclear magnetic resonance according to claim 3, characterized in that: The process of determining the freshness of the fruit according to the color classification and the appearance characteristics is as follows: The stem brightness corresponds to a stem brightness level, the skin gloss corresponds to a gloss level, and the freshness is determined based on the stem brightness level and the gloss level.

7. The fruit tracing method based on low-field magnetic nuclear magnetic resonance according to claim 1, characterized in that: The step of determining the sweetness and acidity of the pulp according to the low-field magnetic resonance signal comprises: Determining the sweetness corresponding to the low-field magnetic resonance signal by using a correlation model that characterizes the low-field magnetic resonance signal per unit mass and the sweetness of the fruit; The acidity corresponding to the low-field magnetic resonance signal is determined by a fitting model that characterizes the transverse relaxation time and the acidity of the fruit.

8. The fruit tracing method based on low-field magnetic nuclear magnetic resonance according to any one of claims 1 to 6, characterized in that: Said elements used to determine the elemental origin of the fruit include potassium, cobalt, rubidium, nickel, manganese, calcium, hydrogen stable isotopes and oxygen stable isotopes; Determining the origin of the element by cluster analysis of the element content; The appearance origin is the continental origin, and the element origin is the national origin; Determine whether the geographical scope of the appearance origin includes the geographical scope of the element origin. If so, the fruit does not have counterfeit origin. Otherwise, the fruit has counterfeit origin.

9. A fruit traceability system based on low-field magnetic nuclear magnetic resonance, characterized in that: The fruit tracing system is equipped with the fruit tracing method based on low-field magnetic nuclear magnetic resonance as claimed in any one of claims 1 to 8, and the fruit tracing system comprises: Vision sensor to obtain images of the fruit; A low-field magnetic resonance analyzer is used to perform low-field magnetic resonance on the fruit core to obtain low-field magnetic resonance signals; Inductively coupled plasma mass spectrometer to obtain the elemental content of fruit cores; A control device communicates with the inductively coupled plasma mass spectrometer and the low-field magnetic resonance analyzer, and is used to determine whether the appearance and appearance origin of the fruit are forged based on the image and the low-field magnetic resonance signal, and to determine whether the origin of the fruit is fake based on the appearance origin and the element content.

10. The fruit traceability system based on low-field magnetic nuclear magnetic resonance according to claim 9, characterized in that: The control device is equipped with a detection system cloud platform to generate a fruit detection report including the image, the low-field magnetic resonance signal, the element content, whether the appearance is forged, and whether the origin is fake.

Images