Peach internal quality detection method and device based on spatial spectrum interception

Abstract

The invention provides a peach internal quality detection method and device based on spatial spectrum interception, and relates to the technical field of nondestructive testing of fruits and vegetables, the method comprises the following steps: collecting multi-point spatial spectrum data of nectarines to be detected; sequentially removing first regional spectral data and second regional spectral data in the multi-point spatial spectral data to obtain effective regional spectral data; carrying out mean value calculation on the effective area spectrum data to obtain target spectrum data; inputting the target spectral data into a sugar degree detection model to obtain a sugar degree detection result of the nectarine to be detected output by the sugar degree detection model; and inputting the target spectral data into a defect detection model to obtain a defect detection result of the nectarine to be detected output by the defect detection model. According to the peach internal quality detection method based on spatial spectrum interception provided by the invention, nondestructive detection of nectarines to be detected is realized, and the detection efficiency and accuracy are improved.

Description

Technical Field

[0001] This invention relates to the field of non-destructive testing technology for fruits and vegetables, and in particular to a method and apparatus for detecting the internal quality of peaches based on spatial spectral interception. Background Technology

[0002] The quality of nectarines is closely related to factors such as sugar content, ripeness, and internal defects. Traditional methods for testing nectarine quality mainly rely on destructive physicochemical analysis, such as using a refractometer to measure soluble solids content to represent sugar content, or cutting open nectarines to identify physiological defects such as pit cracking and pest infestations. These methods are time-consuming, labor-intensive, and destructive to samples, and cannot achieve rapid testing in large batches, failing to meet the urgent need for efficient and non-destructive testing in fruit quality grading, packaging, and sales.

[0003] Existing technologies use transmission near-infrared spectroscopy to construct a predictive model for the internal quality of nectarines and detect the internal quality of the fruit. However, the existing predictive models only collect one integral spectral curve for each detection target, which makes it difficult to accurately characterize the overall soluble solids content of the nectarine and the local defects in the pit, resulting in low detection accuracy. Summary of the Invention

[0004] This invention provides a method and apparatus for detecting the internal quality of peaches based on spatial spectral interception, in order to solve the technical problem of low accuracy and efficiency in the detection of the internal quality of peaches in the prior art.

[0005] This invention provides a method for detecting the internal quality of peaches based on spatial spectral interception, comprising the following steps: Collect multi-point spatial spectral data of the nectarine to be tested; The first region spectral data and the second region spectral data in the multi-point spatial spectral data are removed sequentially to obtain the effective region spectral data; the first region spectral data is the spectral data corresponding to the edge region of the peel of the nectarine to be tested; the second region spectral data is the spectral data corresponding to the pit region of the nectarine to be tested. The mean value of the spectral data of the effective region is calculated to obtain the target spectral data; The target spectral data is input into the sugar content detection model to obtain the sugar content detection result of the nectarine to be tested output by the sugar content detection model; the sugar content detection model is trained by using the target spectral data of various nectarines as feature data and the soluble solids content of the various nectarines as label data. The target spectral data is input into the defect detection model to obtain the defect detection result of the peach to be tested output by the defect detection model; the defect detection model is trained by using the target spectral data of multiple peaches as feature data and the defect categories of the multiple peaches as label data; the defect categories include defective and non-defective.

[0006] According to the present invention, a method for detecting the internal quality of peaches based on spatial spectral interception, wherein the step of sequentially removing the first region spectral data and the second region spectral data from the multi-point spatial spectral data to obtain the effective region spectral data includes: Remove the spectral data of the first region from the multi-point spatial spectral data to obtain the spectral data of the central region; Remove the second region spectral data from the central region spectral data to obtain the effective region spectral data.

[0007] According to the present invention, a method for detecting the internal quality of peaches based on spatial spectral interception, wherein removing the first region spectral data from the multi-point spatial spectral data to obtain the central region spectral data includes: Determine the left edge spectral proportion and the right edge spectral proportion of the multi-point spatial spectral data; The spectral data of the first region are determined based on the spectral ratio of the left edge and the spectral ratio of the right edge. Remove the first region spectral data from the multi-point spatial spectral data to obtain the central region spectral data.

[0008] According to the present invention, a method for detecting the internal quality of peaches based on spatial spectral interception is provided, wherein when the spectral proportion of the left edge is greater than or equal to 50%, the spectral proportion of the right edge is less than 50%. When the spectral proportion of the right edge is greater than or equal to 50%, the spectral proportion of the left edge is less than 50%.

[0009] According to the present invention, a method for detecting the internal quality of peaches based on spatial spectral interception, wherein removing the second region spectral data from the central region spectral data to obtain the effective region spectral data includes: Determine the central spectral curve of the central region spectral data; The left and right spectral curves of the central spectral curve are determined according to a preset ratio. The central spectral curve, the left spectral curve, and the right spectral curve are used as the spectral data of the second region; Remove the second region spectral data from the central region spectral data to obtain the effective region spectral data.

[0010] According to the present invention, a method for detecting the internal quality of peaches based on spatial spectral interception, wherein the step of calculating the mean of the spectral data of the effective region to obtain the target spectral data includes: The mean value of the spectral data of the effective region is calculated to obtain the original calculated spectral data; Based on preset optical reference spectral data, the original calculated spectral data is converted into absorption spectral data; The absorption spectrum data is denoised to obtain the target spectrum data.

[0011] The present invention also provides a device for detecting the internal quality of peaches based on spatial spectral interception, comprising the following modules: The acquisition module is used to acquire multi-point spatial spectral data of the peach to be tested; The removal module is used to sequentially remove the first region spectral data and the second region spectral data from the multi-point spatial spectral data to obtain the effective region spectral data; the first region spectral data is the spectral data corresponding to the edge region of the peel of the nectarine to be tested; the second region spectral data is the spectral data corresponding to the pit region of the nectarine to be tested. The calculation module is used to calculate the mean of the spectral data of the effective region to obtain the target spectral data; The sugar content detection module is used to input the target spectral data into the sugar content detection model and obtain the sugar content detection result of the nectarine to be tested output by the sugar content detection model; the sugar content detection model is trained by using the target spectral data of multiple nectarines as feature data and the soluble solids content of the multiple nectarines as label data. The defect detection module is used to input the target spectral data into the defect detection model and obtain the defect detection result of the peach to be tested output by the defect detection model; the defect detection model is trained by using the target spectral data of multiple peaches as feature data and the defect categories of the multiple peaches as label data; the defect categories include defective and non-defective.

[0012] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the peach internal quality detection method based on spatial spectral interception as described above.

[0013] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for detecting the internal quality of peaches based on spatial spectral interception as described above.

[0014] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the method for detecting the internal quality of peaches based on spatial spectral interception as described above.

[0015] This invention provides a method and apparatus for detecting the internal quality of peaches based on spatial spectral interception. By collecting multi-point spatial spectral data of the peach under test, high-resolution spatially resolved near-infrared transmission spectral data of different regions inside the peach are obtained, avoiding the limitations of a single integrated spectrum and providing a rich data foundation for subsequent analysis. The first and second region spectral data are sequentially removed from the multi-point spatial spectral data to obtain the effective region spectral data. The first region spectral data corresponds to the spectral data of the peel edge region of the peach under test, and the second region spectral data corresponds to the spectral data of the pit region. This process removes noisy and signal-attenuated interference regions (such as the peel edge and pit), extracting the pure pulp region spectrum, reducing irrelevant variations and noise interference, and improving the signal-to-noise ratio of the subsequent model input data. The effective region spectral data is then mean-calculated to obtain the target spectral data, and the effective region is integrated through mean calculation. The system obtains stable spectral curves representing the overall peach fruit by analyzing spectral information, reducing random errors and local variations, and providing consistent and reliable feature data for subsequent model inputs. The target spectral data is then input into a sugar content detection model to obtain the sugar content detection results of the peaches being tested. This model is trained using target spectral data of various peaches as feature data and soluble solids content of various peaches as label data. Similarly, the target spectral data is input into a defect detection model to obtain defect detection results of the peaches being tested. This model is trained using target spectral data of various peaches as feature data and defect categories of various peaches as label data. Defect categories include both defective and non-defective varieties, thus achieving non-destructive and accurate assessment of the sugar content of the peaches being tested, as well as automatic screening of defective or rotten fruit. This meets the needs of online grading of peach quality and improves detection efficiency and accuracy. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0017] Figure 1 This is a flowchart illustrating the method for detecting the internal quality of peaches based on spatial spectral interception provided by the present invention.

[0018] Figure 2This is a schematic diagram of the structure of the short integral full transmission near-infrared spectroscopy online detection device provided by the present invention.

[0019] Figure 3 This is a schematic diagram of the spatially resolved spectrum of the nectarine to be tested provided by the present invention.

[0020] Figure 4 This is a schematic diagram of multi-point spatial spectral data provided by the present invention.

[0021] Figure 5 This is a schematic diagram of the spectral data of the central region provided by the present invention.

[0022] Figure 6 This is a schematic diagram of the effective region spectral data provided by the present invention.

[0023] Figure 7 This is a schematic diagram of the internal quality detection device for peaches based on spatial spectral interception provided by the present invention.

[0024] Figure 8 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0025] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0026] Existing technologies employ transmission near-infrared spectroscopy to construct predictive models for the internal quality of nectarines, enabling the detection of the fruit's internal quality. The principle involves utilizing the absorption of specific frequencies of light by hydrogen-containing groups (such as OH, CH, and NH) in the near-infrared band (780-2500 nm) of organic matter. By measuring the absorption or transmission spectra of the sample and combining this with chemometric methods, a quantitative relationship model is established between spectral information and the quality indicators to be measured, thereby enabling rapid prediction of the quality of unknown samples.

[0027] In nectarine quality testing, near-infrared spectroscopy primarily employs two optical path acquisition modes: diffuse reflectance and transmission. Diffuse reflectance is often used to measure the quality information of the surface or shallow tissues of nectarines. For fruits like nectarines, where the soluble solids content is unevenly distributed, the information carried by diffuse reflectance light has spatial limitations, making it difficult to reflect the overall sugar content of the fruit. Furthermore, it is easily affected by fruit size, shape, surface color, and texture. In addition, diffuse reflectance light is not sensitive enough to quality changes in the core area of ​​the fruit, resulting in lower model prediction accuracy and robustness. Compared to diffuse reflectance, transmission mode acquires transmitted light that penetrates the entire fruit, thus obtaining more comprehensive and in-depth internal information. It can more directly and accurately reflect the overall soluble solids content and internal defects.

[0028] However, in the actual process of constructing a predictive model for the internal quality of nectarines, the presence of the nectarine pit causes severe attenuation of the spectral signal, and the complexity of the internal tissue of the nectarine interferes with the effective information. Therefore, in order to obtain high-quality, high signal-to-noise ratio transmission spectra, the integration time of spectral acquisition is usually increased as much as possible.

[0029] In current common online spectral detection operations, only one integrated spectral curve is typically acquired for each detection target (nectarine). However, integrated spectroscopy can mask local feature differences between different regions of the target, especially for pitted nectarines with significant differences in internal tissue structure. The integrated spectral signal is difficult to truly reflect the specific characteristics of the pulp or pit regions. Therefore, relying on a single integrated spectral curve is insufficient to accurately characterize the overall soluble solids content of the peach and the local defects in the pit, resulting in low accuracy in detecting the internal quality of the peach.

[0030] Therefore, the present invention provides a method and apparatus for detecting the internal quality of peaches based on spatial spectral interception, aiming to improve the accuracy and efficiency of detecting the internal quality of peaches.

[0031] The following is combined Figures 1 to 8 This invention describes a method and apparatus for detecting the internal quality of peaches based on spatial spectral interception.

[0032] Figure 1 This is a flowchart illustrating the method for detecting the internal quality of peaches based on spatial spectral interception provided by the present invention, as shown below. Figure 1 As shown, the method includes the following steps: Step 101: Collect multi-point spatial spectral data of the nectarine to be tested; Specifically, in this embodiment of the invention, a short-integral full-transmission near-infrared spectroscopy online detection device is used to acquire multi-point spatially resolved spectral data of the nectarine to be tested. The nectarine to be tested can be any type of nectarine, such as red nectarine or yellow nectarine.

[0033] Figure 2This is a schematic diagram of the short-integral full-transmission near-infrared spectroscopy online detection device provided by the present invention, as shown below. Figure 2 As shown, the online testing equipment mainly includes: A conveyor belt is used to transport peaches to be tested at a constant speed. Fruit trays are placed on the conveyor belt to hold the target peaches (i.e., the peaches to be tested) and to prevent them from being damaged. A position sensor, placed on the conveyor belt, is used to detect the position of the fruit tray in real time; The light source, cover plate, spectrometer, fan, and light-blocking curtain together constitute the transmission spectrum acquisition optical path, which is used to obtain information on all tissue components along the diameter direction of the cross-section of the peach to be tested; An industrial control computer is used to remotely control the aforementioned online testing equipment via a wireless communication network.

[0034] Specifically, the spectrometer is used to acquire spectral signals with a signal-to-noise ratio higher than a preset threshold within a short integration time of 5 to 10 ms, thereby collecting multi-point spatial spectral data when the peach to be tested passes through the detection area.

[0035] For example, when the nectarine to be tested, placed on the fruit tray, enters the detection area via conveyor belt at a speed of 500 mm / s, the position sensor outputs a trigger signal, and the spectrometer will collect a series of transmission spectrum data in an integration time of 5 ms. The length of the detection target area corresponding to each spectral curve is: 500mm / s × 0.005s = 2.5mm For a peach with an average length of 70 mm, approximately 28 spectra can be obtained in a single sampling using the transmission spectral acquisition optical path, with each spectral line corresponding to a different region of the target.

[0036] Figure 3 This is a schematic diagram of the spatially resolved spectrum of the nectarine to be tested provided by the present invention, as shown below. Figure 3 As shown, based on this series of multi-point spatial spectral data, both a comprehensive overall analysis and a regionally selective analysis of the peaches under test can be performed, thus avoiding the limitations of a single integral spectrum and providing a rich and accurate data foundation for subsequent analysis.

[0037] Step 102: Sequentially remove the first region spectral data and the second region spectral data from the multi-point spatial spectral data to obtain the effective region spectral data; the first region spectral data is the spectral data corresponding to the edge region of the peel of the nectarine to be tested; the second region spectral data is the spectral data corresponding to the pit region of the nectarine to be tested. Specifically, based on the physicochemical properties of nectarines, spectral data from interfering regions are removed at a certain ratio, thereby retaining the effective spectral data. The interfering regions of nectarines mainly include the following two areas: The periphery of the fruit peel: This refers to the area where the peel and pulp meet. In this region, the spectral signal is significantly affected by scattering and reflection from the peel surface, the information content of the internal tissue is too low, and the intensity of irrelevant interference noise is too high or even easily saturates.

[0038] The pit region: In this region, the effective information from the pulp is difficult to penetrate, and the spectral signal is severely attenuated.

[0039] By sequentially removing the first region spectral data corresponding to the peel edge region and the second region spectral data corresponding to the core region from the multi-point spatial spectral data, effective region spectral data is obtained, thereby extracting pure pulp region spectra, reducing irrelevant variations and noise interference, and improving the signal-to-noise ratio of subsequent model input data.

[0040] Step 103: Calculate the mean value of the spectral data of the effective region to obtain the target spectral data; Further, the step of calculating the mean of the spectral data of the effective region to obtain the target spectral data includes: The mean value of the spectral data of the effective region is calculated to obtain the original calculated spectral data; Based on preset optical reference spectral data, the original calculated spectral data is converted into absorption spectral data; The absorption spectrum data is denoised to obtain the target spectrum data.

[0041] Specifically, the average of the spectral data of the extracted effective region is used as the original calculated spectral data of the peach to be tested.

[0042] The transmission spectrum data of a 5mm thick polytetrafluoroethylene plate was used as the optical reference spectrum data. The original calculated spectrum data was corrected using the following expression to obtain the absorption spectrum data. ): The absorption spectral data is denoised to eliminate noise interference, remove baseline drift, and enhance feature information to obtain the target spectral data.

[0043] The denoising process can employ any of the following methods: Savitzky-Golay (SG) smoothing and wavelet transform (WT) to reduce noise interference; Standard Normal Variate (SNV) transform and Multiplicative Scatter Correction (MSC) to eliminate baseline drift; Orthogonal signal correction (OSC) to remove orthogonal variations unrelated to the target; and combinations of SG smoothing with SNV and OSC, respectively.

[0044] This invention integrates the spectral information of the effective region through mean calculation to obtain a stable spectral curve representing the entire peach fruit. Through spectral correction and denoising, environmental interference is reduced and random errors are minimized, providing consistent and reliable target spectral data for subsequent model input.

[0045] Step 104: Input the target spectral data into the sugar content detection model to obtain the sugar content detection result of the nectarine to be tested output by the sugar content detection model; the sugar content detection model is trained using the target spectral data of various nectarines as feature data and the soluble solids content of the various nectarines as label data. Specifically, the sugar content detection model can be constructed using algorithms such as Partial Least Squares Regression (PLSR) or Support Vector Machine (SVM), and then trained under supervision using the target spectral data of various nectarines as feature data and the corresponding soluble solids content as label data.

[0046] In this embodiment of the invention, 189 nectarine samples were selected, including 140 red nectarines and 49 yellow nectarines. To improve the generalization of the sugar content detection model, the selected peaches were of varying sizes and were placed at room temperature for 12 hours before the experiment to reduce the influence of temperature. Multi-point spatially resolved spectral data of the nectarine samples were acquired using a short-integral full-transmission near-infrared spectroscopy online detection device, and the processing steps 102 and 103 described above were performed to obtain the corresponding target spectral data, which served as feature data.

[0047] Peaches were cut open, pitted, and all pulp was juiced. The juice was shaken well and then dripped onto a digital refractometer to measure the soluble solids content. Before each measurement, the probe of the digital refractometer was rinsed with purified water and dried. The average of three measurements was taken as the result of that measurement. All measured values ​​of soluble solids content were used as the physicochemical standard for constructing a quantitative sugar content prediction model, i.e., the label data.

[0048] The dataset was partitioned using a sample set partitioning based on joint XY distances (SPXY), with the training set comprising 80% and the remaining 20% ​​used as the prediction set to evaluate the performance of the model after training.

[0049] For the training set samples, feature data is used as input variables and label data is used as target vectors, which are then input into the sugar content detection model for training. For the prediction set samples, feature data that has undergone the same spectral processing steps as the training set is input into the constructed sugar content detection model to obtain the predicted sugar content dataset. This dataset is then compared with the soluble solids content of the prediction set samples to evaluate the model's performance.

[0050] Using correlation coefficient ( ), root mean square error ( ) and relative analysis error of prediction set ( RPD This is used as an evaluation metric for the model. The calculation method is as follows: In the formula, The number of samples in the training set and the prediction set, respectively. , These are the measured and predicted values ​​of the sample, respectively. for The average value. The root mean square error of the prediction set. The standard deviation of the true values ​​of the prediction set is calculated as follows: In the formula, p The number of samples in the prediction set. , These represent the true and average values ​​of soluble solids content in the prediction set samples, respectively.

[0051] It should be noted that the correlation coefficient of the training set ( R c ) and the correlation coefficient of the test set ( R pThe closer the value is to 1, the better the correlation of the model; the root mean square error of the training set ( ) and root mean square error of the test set ( The closer the values ​​are to each other and the smaller the values ​​are, the higher the prediction accuracy of the model; the relative analysis error of the prediction set RPD The higher the value, the stronger the model's ability to distinguish and judge.

[0052] After obtaining the sugar content detection model through supervised training, the target spectral data is input into the sugar content detection model to obtain the sugar content detection result of the peach to be tested output by the sugar content detection model.

[0053] This invention, through the construction and training of a sugar content detection model, achieves a non-destructive and accurate assessment of the sugar content of peaches under test, meeting the requirements for online grading of peach quality and improving detection efficiency and accuracy.

[0054] Step 105: Input the target spectral data into the defect detection model to obtain the defect detection result of the peach to be tested output by the defect detection model; the defect detection model is trained by using the target spectral data of multiple peaches as feature data and the defect categories of the multiple peaches as label data; the defect categories include defective and non-defective.

[0055] Specifically, the defect detection model can be constructed using algorithms such as Partial Least Squares Discriminant Analysis (PLS-DA) or SVM, and then trained under supervision using target spectral data of various nectarines as feature data and defect categories of various nectarines as label data.

[0056] In this embodiment of the invention, 867 nectarine samples were selected, including 720 red nectarines and 147 yellow nectarines. To improve the generalization of the defect detection model, the selected peaches were of varying sizes and were placed at room temperature for 12 hours before the experiment to reduce the influence of temperature. Multi-point spatially resolved spectral data of the nectarine samples were acquired using a short-integral full-transmission near-infrared spectroscopy online detection device, and the processing steps 102 and 103 described above were performed to obtain the corresponding target spectral data, which served as feature data.

[0057] Internal defects in the area near the pit mainly include physiological defects such as pit splitting and insect damage. In this embodiment of the invention, the peach is cut open to determine whether pit splitting exists and whether insect damage is present near the pit. If either defect is present, the defect category is classified as a positive sample (i.e., defective); otherwise, it is classified as a negative sample (i.e., no defect), thus obtaining label data including both defective and no-defect samples.

[0058] The Kennard-Stone (KS) algorithm was used to split the dataset, with 80% allocated to the training set and the remaining 20% ​​to the prediction set, in order to evaluate the performance of the model after training.

[0059] For the training set samples, feature data are used as input variables, and the defect categories of the training set samples are used as target vectors. These are then input into the defect detection model for training. For the prediction set samples, feature data that has undergone the same spectral processing steps as the training set are input into the constructed defect detection model. The resulting detection results are compared with the true labels of the prediction set samples to evaluate the model's performance.

[0060] use The score, used as an evaluation metric for the defect detection model, is calculated as follows: In the formula, TP This represents the number of correctly predicted positive samples (i.e., the number of correctly picked bad fruits). FP This represents the number of samples that were incorrectly predicted as positive (i.e., good results were misclassified as bad results). FN This represents the number of samples that were incorrectly predicted as negative (i.e., bad results were mistakenly identified as good results).

[0061] The score provides a single, comprehensive standard, forcing defect detection models to find the optimal balance between high precision (few false positives) and high recall (few false negatives). The higher the score, the more effectively the model can detect defective nectarines while minimizing the chance of accidentally damaging undefective ones.

[0062] This invention, through the construction and training of a defect detection model, achieves non-destructive and accurate assessment of physiological defects and pests in peaches under test. This enables automatic screening of defective and rotten fruit, meeting the need for online quality grading of peaches with irregular shapes and significant internal structural differences, and improving detection efficiency and accuracy.

[0063] This invention provides a method for detecting the internal quality of peaches based on spatial spectral interception. By collecting multi-point spatial spectral data of the peach under test, high-resolution spatially resolved near-infrared transmission spectral data of different regions inside the peach are obtained, avoiding the limitations of a single integrated spectrum and providing a rich data foundation for subsequent analysis. The method sequentially removes the first and second region spectral data from the multi-point spatial spectral data to obtain the effective region spectral data. The first region spectral data corresponds to the spectral data of the peel edge region of the peach under test, and the second region spectral data corresponds to the spectral data of the pit region. This process removes noisy and signal-attenuated interference regions (such as the peel edge and pit), extracting the pure pulp region spectrum, reducing irrelevant variations and noise interference, and improving the signal-to-noise ratio of the subsequent model input data. The effective region spectral data is then mean-calculated to obtain the target spectral data, and the effective region is integrated through mean calculation. Spectral information is used to obtain a stable spectral curve representing the overall peach fruit, reducing random errors and local variations, and providing consistent and reliable feature data for subsequent model input. The target spectral data is then input into a sugar content detection model to obtain the sugar content detection results of the peaches being tested. This sugar content detection model is trained using target spectral data of various nectarines as feature data and soluble solids content of various nectarines as label data. Similarly, the target spectral data is input into a defect detection model to obtain the defect detection results of the peaches being tested. This defect detection model is trained using target spectral data of various nectarines as feature data and defect categories of various nectarines as label data. Defect categories include both defective and non-defective, thus achieving non-destructive and accurate assessment of the sugar content of the peaches being tested, as well as automatic screening of defective or rotten fruit, meeting the needs of online peach quality grading and improving detection efficiency and accuracy.

[0064] Further, the step of sequentially removing the first region spectral data and the second region spectral data from the multi-point spatial spectral data to obtain the effective region spectral data includes: Remove the spectral data of the first region from the multi-point spatial spectral data to obtain the spectral data of the central region; Remove the second region spectral data from the central region spectral data to obtain the effective region spectral data.

[0065] Further, removing the first region spectral data from the multi-point spatial spectral data to obtain the central region spectral data includes: Determine the left edge spectral proportion and the right edge spectral proportion of the multi-point spatial spectral data; The spectral data of the first region are determined based on the spectral ratio of the left edge and the spectral ratio of the right edge. Remove the first region spectral data from the multi-point spatial spectral data to obtain the central region spectral data.

[0066] Specifically, Figure 4 This is a schematic diagram of multi-point spatial spectral data provided by the present invention, such as... Figure 4 As shown, the horizontal axis represents wavelength, and the vertical axis represents light intensity. When the peach to be tested passes through the detection area, the short-integral full-transmission near-infrared spectroscopy online detection device acquires... n A dataset consisting of spectral curves (i.e., multi-point spatial spectral data) is denoted as a dataset. S 0={ S 1, S 2, S 3, ... S n}

[0067] make a o The spectral proportion of the left edge. b o Given the spectral proportion of the right edge, the spectral data of the first region includes the number of spectral bars from the left edge. and the number of spectral bars on the right edge They are respectively: In the formula, This indicates rounding down to the nearest integer to ensure that the total number of removed spectra is not zero. n .

[0068] in, a o , b o The range of values ​​for is: 0≤ a o <100, 0≤ b o <100, and a o + b o <100.

[0069] Figure 5 This is a schematic diagram of the central region spectral data provided by the present invention, as shown below. Figure 5 As shown, the horizontal axis represents wavelength, and the vertical axis represents light intensity. This represents the number of spectral curves retained after removing the spectral data from the first region. for: n d1 = nL 1 -R 1 The retained spectral curve dataset (i.e., the spectral data of the central region) for: The embodiments of the present invention achieve adaptive truncation of spectral data through scaling processing, and remove spectral data corresponding to fruit peel edge regions with high noise and signal attenuation, thereby improving data accuracy.

[0070] Furthermore, if the spectral proportion of the left edge is greater than or equal to 50%, the spectral proportion of the right edge is less than 50%. When the spectral proportion of the right edge is greater than or equal to 50%, the spectral proportion of the left edge is less than 50%.

[0071] Specifically, the spectrum of the left (or right) half of the nectarine being tested can be completely removed, retaining only the spectral data of the right (or left) half. In other words, if a o The spectral proportion of the left edge. b o If the spectral proportion is at the right edge, then... a o , b o Each of them can have a value greater than 50, but neither can have a value greater than 50 at the same time.

[0072] For example, in some special cases, if the left (or right) half of the peach being tested is fixedly blocked or interfered with during the test, or if all the peaches pass through the test area in the same posture and orientation, then only the spectral data of the right (or left) half of the area is retained as the spectral data of the central area.

[0073] The embodiments of the present invention flexibly set the spectral ratio of the left edge and the spectral ratio of the right edge to adapt to abnormal situations in actual production lines, thereby ensuring detection continuity and reducing false detection rate.

[0074] Further, removing the second region spectral data from the central region spectral data to obtain the effective region spectral data includes: Determine the central spectral curve of the central region spectral data; The left and right spectral curves of the central spectral curve are determined according to a preset ratio. The central spectral curve, the left spectral curve, and the right spectral curve are used as the spectral data of the second region; Remove the second region spectral data from the central region spectral data to obtain the effective region spectral data.

[0075] Specifically, for the spectral data of the central region: Let its central spectral curve be S c ,but In the formula, This indicates rounding.

[0076] The method for removing the spectral data of the second region (the spectral data corresponding to the kernel region) is as follows: S c Centered on the target, remove each component according to a preset ratio. S c The spectral curves on the left and right, and S c itself.

[0077] make a i , b i If the left and right center spectra are removed at different rates, then the number of spectra on the left side is calculated. and the number of spectral curves on the right They are respectively: In the formula, This indicates rounding down to the nearest integer to ensure that the total number of removed spectra is not zero. .

[0078] in, a i , b i The range of values ​​for is: 0≤ a i <50, 0≤ b i <50.

[0079] Figure 6 This is a schematic diagram of the effective region spectral data provided by the present invention, such as... Figure 6 As shown, the horizontal axis represents wavelength, and the vertical axis represents light intensity. The number of spectral curves retained after removing the spectral data from the second region is also shown. n d2 for: n d2 = n d1 -L 2 -R 2-1 The retained spectral curve dataset (i.e., the effective region spectral data) for: Based on the above embodiments, the embodiments of the present invention experimentally verified the impact of removing the first region spectral data and the second region spectral data from multi-point spatial spectral data according to different spectral cut-off ratios on the sugar content detection model and the defect detection model.

[0080] Table 1 shows the impact of removing spectral data from the first region on the performance of the sugar content detection model provided by this invention. As shown in Table 1, it can be seen that removing the spectral data corresponding to the peel edge region significantly improves the model performance; however, excessive spectral cut-off can lead to the loss of spectral information related to soluble solids, thus affecting model performance. Therefore, it is necessary to carefully select an appropriate removal ratio. In this embodiment of the invention, when the spectral ratio of the left edge... a o =5%, right edge spectral proportion b o The effect is best when the concentration is 5%.

[0081] Table 1. Results of the impact of removing spectral data from the first region on the performance of the sugar content detection model. Table 2 shows the performance impact of removing the second region spectral data on the sugar content detection model provided by the present invention. As shown in Table 2, based on removing the first region spectral data, the second region spectral data is further removed, that is, the central spectral curve, the left spectral curve, and the right spectral curve are removed.

[0082] Will pass through [ a o %, b o The spectral dataset obtained by extracting the spectral data corresponding to the spectral region at the edge of the fruit peel [%]=[5%, 5%]. S 1. Using this as the base data, central spectral interception processing is performed. This includes directly using... S 2. Modeling, and S 2. The case where modeling is performed after spectral preprocessing (referring to the processing in step 103 above). It can be seen that regardless of whether spectral preprocessing is performed, the model trained on the data after removing the spectral data corresponding to the fruit pit region shows improved performance in all aspects. In this embodiment of the invention, when the left center spectral removal ratio... a i =7.5%, Right-side center spectral removal ratio b i When the spectral data is preprocessed using the SG smoothing + SNV transformation combination method in step 103, the model constructed has the best performance when the spectral data is 5% and the model performance is the best.

[0083] Table 2. Results of the impact of removing spectral data from the second region on the performance of the sugar content detection model. Table 3 shows the performance impact of removing the second region spectral data on the defect detection model provided by the present invention. As shown in Table 3, based on removing the first region spectral data, the second region spectral data is further removed, that is, the central spectral curve, the left spectral curve, and the right spectral curve are removed.

[0084] Will pass through [ a o %, b o The spectral dataset obtained by extracting the spectral data corresponding to the spectral region at the edge of the fruit peel [%]=[5%, 5%]. S 1. Using this as the base data, a central spectral cutoff was performed. Table 3 shows the data obtained using this method. S 1 and the dataset after removing the left and right center spectra at the optimal cutoff ratio. S 2. Compare the modeling results separately, including directly using S 2. Modeling, and S 2. The modeling process was performed after spectral preprocessing (referring to the processing in step 103 above), and the PLS-DA and SVM modeling methods were compared. It can be seen that regardless of whether spectral preprocessing was performed or which modeling method was used, the model trained on the spectral data after removing the spectral data corresponding to the fruit pit region showed improved performance in all aspects. In this embodiment of the invention, when the left center spectrum removal ratio... a i =20%, Right-side center spectral removal ratio b i When the efficiency is 15%, and the model constructed using the SG smoothing + OSC combined preprocessing method in step 103 and the SVM modeling method has the best performance.

[0085] Table 3. Results of the impact of removing spectral data from the second region on the performance of the defect detection model. This invention, through centralized truncation, removes spectral data corresponding to the fruit pit region with high noise and signal attenuation, extracts the pure fruit pulp region spectrum, reduces irrelevant variations and noise interference, improves the signal-to-noise ratio of the defect detection model input data, and thus improves detection accuracy.

[0086] The following describes the peach internal quality detection device based on spatial spectral interception provided by the present invention. The peach internal quality detection device based on spatial spectral interception described below can be referred to in correspondence with the peach internal quality detection method based on spatial spectral interception described above.

[0087] Figure 7This is a schematic diagram of the internal quality detection device for peaches based on spatial spectral interception provided by the present invention, as shown below. Figure 7 As shown. This invention provides a peach internal quality detection device based on spatial spectral interception, comprising a data acquisition module 701, a removal module 702, a calculation module 703, a sugar content detection module 704, and a defect detection module 705, wherein: The acquisition module 701 is used to acquire multi-point spatial spectral data of the nectarine to be tested; the removal module 702 is used to sequentially remove the first region spectral data and the second region spectral data from the multi-point spatial spectral data to obtain effective region spectral data; the first region spectral data is the spectral data corresponding to the edge region of the peel of the nectarine to be tested; the second region spectral data is the spectral data corresponding to the pit region of the nectarine to be tested; the calculation module 703 is used to calculate the mean of the effective region spectral data to obtain target spectral data; the sugar content detection module 704 is used to input the target spectral data into the sugar content detection model. The sugar content detection model is used to obtain the sugar content detection result of the nectarine to be tested, output by the sugar content detection model. The sugar content detection model is trained using target spectral data of various nectarines as feature data and soluble solids content of the various nectarines as label data. The defect detection module 705 is used to input the target spectral data into the defect detection model and obtain the defect detection result of the nectarine to be tested, output by the defect detection model. The defect detection model is trained using target spectral data of various nectarines as feature data and defect categories of the various nectarines as label data. The defect categories include defective and non-defective.

[0088] This invention provides a peach internal quality detection device based on spatial spectral interception. It acquires high-resolution spatially resolved near-infrared transmission spectral data of different regions within the peach fruit by collecting multi-point spatial spectral data from the peach under test. This avoids the limitations of a single integrated spectrum and provides a rich data foundation for subsequent analysis. The device sequentially removes the first and second region spectral data from the multi-point spatial spectral data to obtain the effective region spectral data. The first region spectral data corresponds to the spectral data of the peel edge region of the peach under test, while the second region spectral data corresponds to the spectral data of the pit region. This process removes noisy and signal-attenuated interference regions (such as the peel edge and pit) to extract the pure pulp region spectrum, reducing irrelevant variations and noise interference, and improving the signal-to-noise ratio of the subsequent model input data. The effective region spectral data is then mean-calculated to obtain the target spectral data, and the effective region spectral data is integrated through mean calculation. Spectral information is used to obtain a stable spectral curve representing the overall peach fruit, reducing random errors and local variations, and providing consistent and reliable feature data for subsequent model input. The target spectral data is then input into a sugar content detection model to obtain the sugar content detection results of the peaches being tested. This sugar content detection model is trained using target spectral data of various nectarines as feature data and soluble solids content of various nectarines as label data. Similarly, the target spectral data is input into a defect detection model to obtain the defect detection results of the peaches being tested. This defect detection model is trained using target spectral data of various nectarines as feature data and defect categories of various nectarines as label data. Defect categories include both defective and non-defective, thus achieving non-destructive and accurate assessment of the sugar content of the peaches being tested, as well as automatic screening of defective or rotten fruit, meeting the needs of online peach quality grading and improving detection efficiency and accuracy.

[0089] Figure 8 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 8 As shown, the electronic device may include: a processor 810, a communications interface 820, a memory 830, and a communication bus 840, wherein the processor 810, the communications interface 820, and the memory 830 communicate with each other via the communication bus 840. The processor 810 can call logical instructions in the memory 830 to execute a method for detecting the internal quality of peaches based on spatial spectral interception. This method includes: Collect multi-point spatial spectral data of the nectarine to be tested; The first region spectral data and the second region spectral data in the multi-point spatial spectral data are removed sequentially to obtain the effective region spectral data; the first region spectral data is the spectral data corresponding to the edge region of the peel of the nectarine to be tested; the second region spectral data is the spectral data corresponding to the pit region of the nectarine to be tested. The mean value of the spectral data of the effective region is calculated to obtain the target spectral data; The target spectral data is input into the sugar content detection model to obtain the sugar content detection result of the nectarine to be tested output by the sugar content detection model; the sugar content detection model is trained by using the target spectral data of various nectarines as feature data and the soluble solids content of the various nectarines as label data. The target spectral data is input into the defect detection model to obtain the defect detection result of the peach to be tested output by the defect detection model; the defect detection model is trained by using the target spectral data of multiple peaches as feature data and the defect categories of the multiple peaches as label data; the defect categories include defective and non-defective.

[0090] Furthermore, the logical instructions in the aforementioned memory 830 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0091] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the peach internal quality detection method based on spatial spectral interception provided by the above methods, the method comprising: Collect multi-point spatial spectral data of the nectarine to be tested; The first region spectral data and the second region spectral data in the multi-point spatial spectral data are removed sequentially to obtain the effective region spectral data; the first region spectral data is the spectral data corresponding to the edge region of the peel of the nectarine to be tested; the second region spectral data is the spectral data corresponding to the pit region of the nectarine to be tested. The mean value of the spectral data of the effective region is calculated to obtain the target spectral data; The target spectral data is input into the sugar content detection model to obtain the sugar content detection result of the nectarine to be tested output by the sugar content detection model; the sugar content detection model is trained by using the target spectral data of various nectarines as feature data and the soluble solids content of the various nectarines as label data. The target spectral data is input into the defect detection model to obtain the defect detection result of the peach to be tested output by the defect detection model; the defect detection model is trained by using the target spectral data of multiple peaches as feature data and the defect categories of the multiple peaches as label data; the defect categories include defective and non-defective.

[0092] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described methods for detecting the internal quality of peaches based on spatial spectral interception, the method comprising: Collect multi-point spatial spectral data of the nectarine to be tested; The first region spectral data and the second region spectral data in the multi-point spatial spectral data are removed sequentially to obtain the effective region spectral data; the first region spectral data is the spectral data corresponding to the edge region of the peel of the nectarine to be tested; the second region spectral data is the spectral data corresponding to the pit region of the nectarine to be tested. The mean value of the spectral data of the effective region is calculated to obtain the target spectral data; The target spectral data is input into the sugar content detection model to obtain the sugar content detection result of the nectarine to be tested output by the sugar content detection model; the sugar content detection model is trained by using the target spectral data of various nectarines as feature data and the soluble solids content of the various nectarines as label data. The target spectral data is input into the defect detection model to obtain the defect detection result of the peach to be tested output by the defect detection model; the defect detection model is trained by using the target spectral data of multiple peaches as feature data and the defect categories of the multiple peaches as label data; the defect categories include defective and non-defective.

[0093] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0094] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0095] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0096] In this application's embodiments, "determine B based on A" means that factor A must be considered when determining B. It is not limited to "B can be determined based solely on A," but should also include: "determine B based on A and C," "determine B based on A, C, and E," "determine C based on A, and further determine B based on C," etc. Additionally, it can include using A as a condition for determining B, for example, "when A meets the first condition, determine B using the first method"; another example, "when A meets the second condition, determine B," etc.; another example, "when A meets the third condition, determine B based on the first parameter," etc. Of course, it can also be a condition where A is a factor in determining B, for example, "when A meets the first condition, determine C using the first method, and further determine B based on C," etc.

[0097] It should also be noted that the terms "target," "first," and "second" in this invention are used to distinguish similar objects, not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, without limiting the number of objects; for example, the first object can be one or more.

[0098] In this invention, the term "multiple" refers to two or more kinds, and other quantifiers are similar.

[0099] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting the internal quality of peaches based on spatial spectral interception, characterized in that, include: Collect multi-point spatial spectral data of the nectarine to be tested; The first region spectral data and the second region spectral data in the multi-point spatial spectral data are removed sequentially to obtain the effective region spectral data; the first region spectral data is the spectral data corresponding to the edge region of the peel of the nectarine to be tested; the second region spectral data is the spectral data corresponding to the pit region of the nectarine to be tested. The mean value of the spectral data of the effective region is calculated to obtain the target spectral data; The target spectral data is input into the sugar content detection model to obtain the sugar content detection result of the nectarine to be tested output by the sugar content detection model; the sugar content detection model is trained by using the target spectral data of various nectarines as feature data and the soluble solids content of the various nectarines as label data. The target spectral data is input into the defect detection model to obtain the defect detection results of the peach to be tested, output by the defect detection model. The defect detection model is trained using the target spectral data of various nectarines as feature data and the defect categories of the various nectarines as label data; the defect categories include defective and non-defective.

2. The method for detecting the internal quality of peaches based on spatial spectral interception according to claim 1, characterized in that, The step of sequentially removing the first region spectral data and the second region spectral data from the multi-point spatial spectral data to obtain the effective region spectral data includes: Remove the spectral data of the first region from the multi-point spatial spectral data to obtain the spectral data of the central region; Remove the second region spectral data from the central region spectral data to obtain the effective region spectral data.

3. The method for detecting the internal quality of peaches based on spatial spectral interception according to claim 2, characterized in that, The step of removing the first region spectral data from the multi-point spatial spectral data to obtain the central region spectral data includes: Determine the left edge spectral proportion and the right edge spectral proportion of the multi-point spatial spectral data; The spectral data of the first region are determined based on the spectral ratio of the left edge and the spectral ratio of the right edge. Remove the first region spectral data from the multi-point spatial spectral data to obtain the central region spectral data.

4. The method for detecting the internal quality of peaches based on spatial spectral interception according to claim 3, characterized in that, When the spectral proportion of the left edge is greater than or equal to 50%, the spectral proportion of the right edge is less than 50%. When the spectral proportion of the right edge is greater than or equal to 50%, the spectral proportion of the left edge is less than 50%.

5. The method for detecting the internal quality of peaches based on spatial spectral interception according to claim 3, characterized in that, The step of removing the second region spectral data from the central region spectral data to obtain the effective region spectral data includes: Determine the central spectral curve of the central region spectral data; The left and right spectral curves of the central spectral curve are determined according to a preset ratio. The central spectral curve, the left spectral curve, and the right spectral curve are used as the spectral data of the second region; Remove the second region spectral data from the central region spectral data to obtain the effective region spectral data.

6. The method for detecting the internal quality of peaches based on spatial spectral interception according to claim 1, characterized in that, The step of calculating the mean of the spectral data of the effective region to obtain the target spectral data includes: The mean value of the spectral data of the effective region is calculated to obtain the original calculated spectral data; Based on preset optical reference spectral data, the original calculated spectral data is converted into absorption spectral data; The absorption spectrum data is denoised to obtain the target spectrum data.

7. A device for detecting the internal quality of peaches based on spatial spectral interception, characterized in that, include: The acquisition module is used to acquire multi-point spatial spectral data of the peach to be tested; The removal module is used to sequentially remove the first region spectral data and the second region spectral data from the multi-point spatial spectral data to obtain the effective region spectral data; the first region spectral data is the spectral data corresponding to the edge region of the peel of the nectarine to be tested; the second region spectral data is the spectral data corresponding to the pit region of the nectarine to be tested. The calculation module is used to calculate the mean of the spectral data of the effective region to obtain the target spectral data; The sugar content detection module is used to input the target spectral data into the sugar content detection model and obtain the sugar content detection result of the nectarine to be tested output by the sugar content detection model; the sugar content detection model is trained by using the target spectral data of multiple nectarines as feature data and the soluble solids content of the multiple nectarines as label data. The defect detection module is used to input the target spectral data into the defect detection model and obtain the defect detection results of the peach to be tested output by the defect detection model. The defect detection model is trained using the target spectral data of various nectarines as feature data and the defect categories of the various nectarines as label data; the defect categories include defective and non-defective.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the method for detecting the internal quality of peaches based on spatial spectral interception as described in any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for detecting the internal quality of peaches based on spatial spectral interception as described in any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for detecting the internal quality of peaches based on spatial spectral interception as described in any one of claims 1 to 6.

Images