Rapeseed near infrared spectrum detection device and method

Abstract

This invention relates to the field of near-infrared spectroscopy detection technology, and discloses a near-infrared spectroscopy detection device and method for rapeseed. The sample-carrying component includes a sample tray and a stirring element. The relative rotation of the sample tray and the stirring element allows the sample in the sample tray to be turned over, achieving the interchange of upper and lower layers and uniform mixing of inner and outer rings. This ensures reliable test results and stable experimental repeatability, reflecting the true data of particulate matter to a greater extent. The central axis of the receiving module forms a certain angle with the optical axis of the light source module, effectively avoiding interference from specular reflection and achieving better overlap between the irradiation spot and the receiving spot. Furthermore, this invention employs a three-level anomaly detection mechanism—spectral integrity, correlation, and dispersion—to eliminate abnormal spectra from multiple acquired spectral data. This effectively eliminates outlier data caused by sample state fluctuations, sample loading differences, or accidental interference, improving the consistency and reliability of the spectral data.

Description

Technical Field

[0001] This invention relates to the field of near-infrared spectroscopy detection technology, and in particular to a near-infrared spectroscopy detection device and method for rapeseed. Background Technology

[0002] Near-infrared spectroscopy, as a rapid and non-destructive detection method, has been widely used in the field of oilseed crop quality testing. It can be used to determine the content of nutrients such as moisture, crude fat (oil content), erucic acid, and oleic acid in rapeseed, providing important basis for rapeseed purchase pricing, oil processing, and quality control.

[0003] Existing near-infrared spectroscopy (NIRS) equipment for rapeseed typically places the rapeseed sample in a sample cup, and the light emitted from a light source illuminates the sample cup, acquiring the spectral information of a ring of the sample surface. The calculated content of moisture, crude fat (oil content), erucic acid, oleic acid, etc., is then used as the overall rapeseed material content. However, in existing equipment, the sample is stationary in the sample cup, and the detection beam only illuminates the surface layer of the sample. This detection method has several significant drawbacks when detecting non-uniform samples such as rapeseed: First, insufficient sample homogeneity leads to poor spectral representativeness. Generally speaking, the better the sample homogeneity, the better the near-infrared model constructed. Rapeseed is a natural product, and the moisture, oil content, and other components vary greatly among different seeds. Furthermore, rapeseed seeds differ in shape, size, and plumpness, and the arrangement of seeds and pore distribution are random during natural sample loading. When the detection beam only illuminates a fixed area on the surface of the sample cup, the collected spectral information comes only from a limited number of surface seeds, which cannot represent the average quality of the entire batch of samples. Inhomogeneous seed size can also cause spectral baseline shifts or changes in characteristic peaks, affecting the accuracy and robustness of the quantitative model.

[0004] Second, the detection beam has limited penetration depth, leading to the loss of information about the sample's interior. Light penetration between rapeseed grains is restricted, making it difficult to obtain spectral information from the middle and bottom layers. Studies of heterogeneous samples show significant differences in the spectra at different test points of the same sample. Collecting spectra from only a single point or a limited area on the surface will miss a large amount of valuable information, resulting in systematic bias in the detection results.

[0005] Third, inconsistent sample loading conditions lead to spectral measurement errors. Variations in sample density during loading significantly affect the results of diffuse reflectance spectral measurements. In multiple tests of the same batch of rapeseed, the surface smoothness and tightness of the loaded sample may differ each time, resulting in spectral differences and fluctuations in predicted values. Although preprocessing methods such as vector normalization can partially reduce the errors caused by sample sparsity, this is a passive compensation and cannot fundamentally eliminate the measurement uncertainties introduced by differences in sample loading conditions.

[0006] To overcome the above problems, existing technology proposes a scheme of repeated testing by repeatedly emptying and refilling the sample cup. After one test, the sample is completely emptied from the sample cup, refilled, leveled, and the spectrum is collected again. This process is repeated several times, and the average of the multiple test results is taken as the final result. The purpose of this scheme is to expand the sampling range by changing the arrangement of the particles during each sample loading, allowing the detection beam to illuminate different seeds, thereby alleviating the problem of insufficient representativeness in a single test to some extent. However, this method still has the following problems: First, it cannot quickly and exhaustively test a sufficient number of samples, thus limiting the representativeness of the test.

[0007] Secondly, the operation is extremely inefficient, time-consuming, and labor-intensive. Each time the sample is poured out and refilled, it is necessary to manually or with the aid of tools completely transfer the sample from the sample cup, then refill and smooth the surface, a cumbersome process. Assuming 5-10 repeated tests, the testing time for a single sample increases exponentially, severely reducing throughput and failing to meet the needs of rapid screening of large batches of samples.

[0008] Third, the uncontrollable sample loading state introduces new sources of error. Each time the sample is poured out and reloaded, it is impossible to guarantee that the loading force, leveling method, and final sample density will be completely consistent with the previous one. This inconsistency in sample loading state itself introduces new spectral differences, leading to increased dispersion of multiple measurement results. It may even lower the accuracy of the average value due to a significant deviation in the loading of a particular sample, thereby reducing the reproducibility of the detection.

[0009] Fourth, the risk of sample exposure increases, making it easier to introduce contamination or cause loss. During repeated pouring and loading, rapeseed seeds are exposed to an open environment for extended periods, easily absorbing moisture and dust from the air, or becoming cross-contaminated with containers and tools, affecting the accuracy of moisture and other indicators. Simultaneously, repeated pouring can cause seed breakage and shell shedding; changes in physical morphology directly affect spectral characteristics. Furthermore, repeated pouring and loading inevitably result in the scattering and loss of a small amount of sample, affecting the representativeness of the results.

[0010] Fifth, it is difficult to integrate with automated testing processes. The solution of repeatedly emptying and reloading relies on manual operation or complex mechanical transfer mechanisms. The structure is complex and the operation cycle is slow, making it difficult to seamlessly connect with high-throughput online or laboratory automated testing systems, which limits its application in large-scale automated testing scenarios. Summary of the Invention

[0011] The purpose of this invention is to provide a near-infrared spectroscopy detection device and method for rapeseed with high detection efficiency and high reliability.

[0012] To achieve the above objectives, the present invention provides a near-infrared spectroscopy detection device for rapeseed, comprising: The light source module is used to provide detection light to illuminate the sample; The calibration module is used for wavelength calibration and background calibration; A sample carrying assembly includes a sample tray and a stirring element, wherein the sample tray is used to hold the sample and the stirring element is disposed inside the sample tray; A rotating sample stage is used to drive the sample disk to rotate relative to the stirring component, so that the sample disk drives the sample inside to tumble and mix in the axial direction and generate convection exchange between the center and the edge in the radial direction during the rotation process, and the rotation center axis of the sample disk is offset from the optical axis of the light source module. The receiving module is used to receive the light signal generated after the sample is irradiated by the detection light; A spectrometer is used to receive the optical signal transmitted by the receiving module and generate corresponding spectral data; A data processing and analysis module is used to process and analyze the spectral data.

[0013] In addition, the present invention also provides a method for near-infrared spectroscopy detection of rapeseed, comprising: The sample to be tested is placed in the sample pan. The sample pan is rotated relative to the stirring component by rotating the sample stage. During the rotation, the sample pan causes the sample inside to tumble and mix in the axial direction and generate convection exchange between the center and the edge in the radial direction. During the rotation of the sample disk, the detection light emitted by the light source module illuminates the sample on the sample disk, and the spectrometer collects the light signal generated by the sample after being illuminated by the detection light, which is received by the receiving module at a preset sampling frequency, to obtain multiple raw spectral data. The calibration module is placed in the optical path of the light source module, and wavelength calibration data and background calibration data are collected by the spectrometer. The data processing and analysis module converts the original spectral data based on the wavelength calibration data and the background calibration data to obtain converted spectral data. Abnormal spectra are removed from the converted spectral data to obtain valid spectral data. The valid spectral data is preprocessed to obtain preprocessed sample spectral data. Then, based on a preset algorithm, the preprocessed sample spectral data is calculated and the corresponding component content value is output.

[0014] Compared with the prior art, the beneficial effects of the present invention are as follows: The sample-carrying assembly of this invention comprises a sample tray and a stirring element. The relative rotation of the sample tray and the stirring element allows for the agitation of the sample within the tray, enabling the exchange of upper and lower layers and uniform mixing of the inner and outer rings. This redistribution of particles at different layers and positions within the sample cup, without altering the overall sample loading state, allows the detection beam to quickly and continuously scan a wider sample area. This effectively balances the requirements of information representativeness, operational efficiency, and detection repeatability, ensuring reliable test results and stable experimental repeatability, and reflecting the true data of particulate matter to a greater extent. Furthermore, this invention utilizes a rotating sample stage to rotate the sample tray while the stirring element remains stationary. The rotation axis of the sample tray is offset from the optical axis of the light source module, preventing the spectrometer from acquiring signals from the stirring element that could affect data integrity, thus improving detection accuracy and reliability. Additionally, the central axis of the receiving module forms a certain angle with the optical axis of the light source module, effectively avoiding interference from specular reflection and achieving better overlap between the irradiated and received light spots, ensuring a sufficiently large spot area. Furthermore, this invention employs a three-level anomaly detection mechanism—spectral integrity, correlation, and dispersion—to eliminate abnormal spectra from multiple collected spectral data. This effectively removes outlier data caused by sample state fluctuations, sample loading differences, or accidental interference, improving the consistency and reliability of spectral data and thus enhancing the accuracy and robustness of subsequent quantitative or qualitative analysis models. Attached Figure Description

[0015] Figure 1 This is a first-view structural schematic diagram of the rapeseed near-infrared spectroscopy detection device according to an embodiment of the present invention; Figure 2 This is a second-view structural schematic diagram of the rapeseed near-infrared spectroscopy detection device according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the sample carrier assembly according to an embodiment of the present invention; Figure 4 This is an exploded view of the sample carrier assembly according to an embodiment of the present invention; Figure 5 This is a cross-sectional view of the sample carrier assembly according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the structure of the rotating sample stage according to an embodiment of the present invention; Figure 7 This is a top view of the rotating sample stage according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the structure of the sample carrier assembly mounted on the rotating sample stage according to an embodiment of the present invention; Figure 9 This is a front view of the stirring component according to an embodiment of the present invention; Figure 10 This is a rear view of the stirring component according to an embodiment of the present invention; Figure 11 This is a left view of the stirring component according to an embodiment of the present invention; Figure 12 This is a right view of the stirring component according to an embodiment of the present invention; Figure 13 This is a top view of the stirring component according to an embodiment of the present invention; Figure 14 This is a bottom view of the stirring component according to an embodiment of the present invention; Figure 15 This is a cross-sectional view of the stirring component according to an embodiment of the present invention; Figure 16 This is a longitudinal cross-sectional view of the stirring component according to an embodiment of the present invention; Figure 17 This is a cross-sectional view of the sample tray according to an embodiment of the present invention; Figure 18 This is a bottom view of the sample tray according to an embodiment of the present invention; Figure 19 These are experimental results diagrams of an embodiment of the present invention; Figure 20 This is a spectrum of rapeseed from an embodiment of the present invention.

[0016] In the diagram, 100-light source module; 110-light source; 120-condenser cup; 130-first lens; 140-extinction lens tube; 200-calibration module; 210-switching device; 300-sample carrier assembly; 310-sample tray; 311-base plate; 312-annular sidewall; 313-retaining ring; 314-driven stop block; 320-stirring component; 321-stirring top surface; 322-first stirring side surface; 323-stirring bottom surface; 324-through groove; 325-second stirring side surface; 330-connector; 331-connecting groove; 332-connecting shaft; 340-bearing; 350-upper support component; 351- 352-Second upper support; 353-First gasket; 360-Lower support; 361-First lower support; 362-Second lower support; 363-Second gasket; 370-Sealing ring; 380-First limiting member; 400-Rotating sample stage; 410-Second limiting member; 420-Stage; 421-Boss; 430-Rotating ring; 431-Drive pawl; 500-Receiving module; 600-Spectrometer; 700-Drive device; 710-Motor; 720-Drive gear; 730-Driven gear; 800-Identification module; 810-Detection device; 820-Identification device. Detailed Implementation

[0017] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0018] In the description of this invention, it should be noted that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0019] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0020] Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0021] Example 1 like Figures 1 to 18 As shown, a preferred embodiment of the present invention provides a rapeseed near-infrared spectroscopy detection device, comprising: The light source module 100 is used to provide detection light to illuminate the sample; Calibration module 200 is used for wavelength calibration and background calibration; The sample carrying assembly 300 includes a sample tray 310 and a stirring element 320. The sample tray 310 is used to hold the sample, and the stirring element 320 is disposed inside the sample tray 310. The rotating sample stage 400 is used to drive the sample disk 310 to rotate relative to the stirring component 320, so that the sample disk 310 drives the sample inside to tumble and mix in the axial direction and generate convection exchange between the center and the edge in the radial direction during the rotation process, and the rotation center axis of the sample disk 310 is offset from the optical axis of the light source module 100. The receiving module 500 is used to receive the light signal generated by the sample after it is irradiated by the detection light. The spectrometer 600 is used to receive the optical signal transmitted by the receiving module 500 and generate the corresponding spectral data; The data processing and analysis module is used to process and analyze spectral data.

[0022] This embodiment sets up a sample tray 310 and a stirring element 320 in the sample carrier assembly 300. By rotating the sample tray 310 and the stirring element 320 relative to each other, the sample in the sample tray 310 can be turned over, realizing the exchange of the upper and lower layers of the sample and the uniform mixing of the inner and outer rings. Without changing the overall filling state of the sample, the seeds in different layers and positions in the sample cup can be redistributed, so that the detection beam can quickly and continuously scan a wider sample area. It can effectively take into account the requirements of information representativeness, operational efficiency and detection repeatability, so as to ensure the reliability of the sampling test results and the stability of experimental repeatability, and reflect the true data of particulate matter to a greater extent.

[0023] like Figure 2 As shown, the light source module 100 of this embodiment includes a light source 110, a condenser cup 120, a first lens 130, and an extinction mirror tube 140. The light source 110, condenser cup 120, and first lens 130 are located in the extinction mirror tube 140 and arranged sequentially along the length of the extinction mirror tube 140. The light emitted from the light source 110 is focused by the condenser cup 120 and then acts on the first lens 130. The first lens 130 converts the light into a vertical straight line and then illuminates the sample in the sample tray 310. The condenser cup 120 is used to converge the light emitted from the light source 110, and the first lens 130 is used to collimate the light. Optionally, the first lens 130 is a plano-convex lens.

[0024] The condenser cup 120 has a concave reflective cavity, and the light source 110 is located at the bottom center of the reflective cavity. The inner wall of the reflective cavity is a paraboloid or ellipsoidal surface for reflecting and converging the detection beam emitted by the light source 110 to the detection area of ​​the sample plate 310. The opening of the condenser cup 120 is set towards the sample plate 310.

[0025] For the extinction lens barrel 140, an extinction lens barrel 140 with an extinction structure and / or an extinction material can be adopted. In terms of structure, the extinction structure refers to the geometric structure set in the inner wall of the lens barrel or the internal optical path to suppress the reflection and propagation of stray light. It reduces the possibility of stray light reaching the imaging surface by changing the incident angle of light, increasing the optical path blockage, or extending the path of non-imaging light rays. For example, the matting structure may include, but is not limited to, inner wall grooves, light-blocking rings, multi-level steps, or rough surfaces formed by sandblasting. In terms of materials, matting materials refer to materials with low reflectivity or high light absorption properties used to form the lens barrel body or its surface treatment layer, which suppress stray light by reducing the reflection of light on the inner wall of the lens barrel. Such materials may include, but are not limited to: coating a black matte coating (such as black chrome, black nickel, carbon nanotube coating, or Acktar metallic black film) on the surface of a metal substrate (such as aluminum alloy or stainless steel); directly injection molding using engineering plastics with high light absorption properties (such as polyether ether ketone PEEK or polysulfone PSU with added carbon black); or forming a low reflectivity surface layer on metal and plastic substrates through processes such as anodizing, chemical deposition, and physical vapor deposition (PVD); or spraying with commercially available high-absorption matting paints (such as Nextel Velvet Coating, Aeroglaze Z306, etc.).

[0026] The rapeseed near-infrared spectroscopy detection device of this embodiment includes a device housing, and a light source module 100, a calibration module 200, a sample carrier component 300, a rotating sample stage 400, a receiving module 500 and a spectrometer 600 are all mounted on the device housing.

[0027] The calibration module 200 includes a first calibration plate, a second calibration plate, and a switching device 210. Both the first and second calibration plates are connected to the switching device 210. The switching device 210 is used to move either the first or second calibration plate into or out of the optical path through which the light source 110 illuminates the sample. The first calibration plate is used for wavelength calibration, and the second calibration plate is used for background calibration. Optionally, the switching device 210 moves either the first or second calibration plate into or out of the optical path through which the light source 110 illuminates the sample by rotation. The first calibration plate has at least three distinguishable characteristic absorption peaks in the near-infrared mid-wave band, and the wavelength positions of each characteristic absorption peak are known and stable. In this embodiment, the first calibration plate is made of a material such as polystyrene that has characteristic absorption in the near-infrared mid-wave band. The second calibration plate has uniform and stable diffuse reflection characteristics in the near-infrared band. In this embodiment, the second calibration plate is a white plate.

[0028] Furthermore, the central axis of the receiving module 500 is set at a certain angle to the optical axis of the light source module 100, so that the incident light path of the light source 110 and the signal receiving light path are set non-coaxially at a preset angle. This effectively avoids interference from specular reflection light, achieves better overlap between the illumination spot and the receiving spot, and ensures a sufficiently large spot area. First, in terms of effectively suppressing specular reflection interference, since the collection direction of the receiving light path deviates from the main emission direction of specular reflection on the sample surface, most of the surface reflection light that does not carry internal chemical information of the sample is physically isolated, significantly improving the signal-to-noise ratio and spectral stability of the diffuse reflection signal, which is especially suitable for samples with a certain gloss on the surface of particles such as rapeseed. Second, it ensures the overlap between the effective detection field of view and the illumination area. Unlike simple tilted illumination, this embodiment precisely matches the angle with the spatial position of the receiving fiber, so that the central area of ​​the field of view of the receiving fiber highly overlaps with the central area of ​​the spot formed by the light source 110 on the sample surface. While effectively filtering specular reflection light, it maximizes the collection efficiency of deep diffuse reflection light. Finally, the angled setting method balances the light spot area and signal strength within a compact space, avoiding the problem of physical obstruction of incident light by the receiving component in the coaxial optical path, thus ensuring a sufficiently large effective illumination area. At the same time, the high overlap between the field of view and the light spot also ensures that a stronger effective signal can be obtained under the same light source power of 110, providing a high signal-to-noise ratio data foundation for subsequent quantitative spectral analysis.

[0029] In this embodiment, the receiving module 500 uses optical fiber, and the central axis of the optical fiber is set at a certain angle to the optical axis of the light source module 100. Optionally, the angle between the central axis of the receiving module 500 and the optical axis of the light source module 100 is 30° to 60°, preferably 45° to 50°. In this embodiment, the light emitted by the light source module 100 is incident perpendicularly on the sample, and the receiving module 500 is set at an angle. In addition, the diameter of the light spot irradiated by the light source module 100 on the sample is 30 to 50 mm, preferably 40 mm, and the edge light intensity uniformity of the light spot should reach more than 90%.

[0030] Example 2 The difference between this embodiment and Embodiment 1 is that, based on Embodiment 1, this embodiment provides further explanation of the sample carrier assembly 300 and the rotating sample stage 400.

[0031] In this embodiment, the top surface of the stirring element 320 is an inclined surface, and the bottom surface of the stirring element 320 is provided with a through groove 324. The through groove 324 extends obliquely from one end of the stirring element 320 away from the rotation center axis of the sample pan 310 towards the end closer to the rotation center axis of the sample pan 310, or the through groove 324 extends obliquely from one end of the stirring element 320 near the rotation center axis of the sample pan 310 towards the end away from the rotation center axis of the sample pan 310. In this embodiment, the through groove 324 extends obliquely from one end of the stirring element 320 away from the rotation center axis of the sample pan 310 towards the end closer to the rotation center axis of the sample pan 310, that is, the end of the through groove 324 away from the rotation center axis of the sample pan 310 is located on the top surface of the stirring element 320, and the end closer to the rotation center axis of the sample pan 310 is located on the side surface of the stirring element 320. When the through groove 324 extends at an angle from the end of the stirrer 320 near the rotation center axis of the sample plate 310 to the end away from the rotation center axis of the sample plate 310, the end of the through groove 324 near the rotation center axis of the sample plate 310 is located on the top surface of the stirrer 320, and the end away from the rotation center axis of the sample plate 310 is located on the side surface of the stirrer 320. During use, the stirring element 320 remains stationary while the sample tray 310 rotates, causing the sample to move towards the stirring element 320. Since the top surface of the stirring element 320 is inclined, it scoops up the sample from the bottom, allowing it to move upwards along the inclined surface. Upon reaching the highest point of the stirring element 320, the sample randomly falls down, landing behind it. Simultaneously, because the stirring element 320 has a through-groove 324 that extends inclined from the outside in, samples located on the outer ring move towards the inner ring during the rotation of the sample tray 310, achieving an exchange between the inner and outer rings. This results in more uniform mixing, collecting more data from the same batch of samples, thus improving detection accuracy. It ensures that all grains in the entire sample cup have the opportunity to participate in the detection process, resulting in highly representative test results.

[0032] Optionally, the through groove 324 can also extend from one end near the rotation center axis of the sample disk 310 to the other end away from the rotation center axis of the sample disk 310. The sample located in the inner ring will move to the outer ring through the through groove 324, realizing the interchange of the inner and outer rings, making the stirring more uniform.

[0033] The stirring component 320 in this embodiment includes a stirring top surface 321, a first stirring side surface 322, a second stirring side surface 325, and a stirring bottom surface 323 connected to each other. The stirring top surface 321 is an inclined surface, and the stirring bottom surface 323 has a through groove 324, with its two ends located on the stirring top surface 321 and the first stirring side surface 322, respectively. In this embodiment, the stirring top surface 321 is an arc surface. Specifically, the radius of the stirring top surface 321 is 75 mm. Furthermore, the angle between the stirring top surface 321 and the stirring bottom surface 323 is 120° to 150°, and in this embodiment, it is 135°. Additionally, in this embodiment, the first stirring side surface 322 is perpendicular to the stirring bottom surface 323. When in use, the sample tray 310 is placed horizontally, so the tangent of the arc surface of the stirring top surface 321 forms an angle greater than 90° with the horizontal plane, making this angle an obtuse angle. This angle mainly facilitates the movement of granular objects such as rapeseed from the bottom layer along the curve of the stirring top surface 321 from bottom to top, and then randomly falls down after reaching the highest point of the stirring element 320. If this angle is too large, it will cause the stirring element 320 to be too thin, resulting in deformation of the stirring element 320. If this angle is too small, the bottom layer of granular grains will not be able to move along the curve of the stirring element 320, resulting in a small portion of the stirring element 320 and the corner of the sample tray 310 always not being mixed. At the same time, the contact area between the bottom of the stirring top surface 321 and the horizontal plane should be as small as possible to prevent granular objects such as rapeseed from getting stuck at the bottom. The second stirring side 325 is an arc surface, and the diameter of the second stirring side 325 is slightly smaller than the inner diameter of the sample pan 310. In this embodiment, the diameter of the second stirring side 325 is 0.1 to 1 mm smaller than the inner diameter of the sample pan 310. During the rotation of the sample pan 310, the sample can be prevented from passing between the second stirring side 325 and the inner wall of the sample pan 310.

[0034] Furthermore, the bottom of the through groove 324 is curved, which reduces the friction when the sample moves within the through groove 324 and prevents the sample from getting stuck. The width of the through groove 324 is greater than 5 mm and less than 12 mm. This prevents the through groove 324 from being too narrow, which would result in the sample moving too slowly between the inner and outer rings of the sample tray 310 and risking it getting stuck in the through groove 324. It also prevents the through groove 324 from being too wide, which would result in the inner and outer rings being positioned too narrowly. For example, a sample that was originally close to the outer ring would still move towards the outer ring after passing through the through groove 324, instead of moving directly to the inner ring.

[0035] In this embodiment, the sample carrying assembly 300 further includes a connector 330, a bearing 340, an upper support 350, and a lower support 360. A mounting hole is provided at the center of the sample tray 310, and the bearing 340 is disposed in the mounting hole. The stirring element 320 is connected to the connector 330, and the connector 330 is fixed relative to the inner ring of the bearing 340. The upper support 350 and the lower support 360 are respectively disposed on both sides of the bottom surface of the sample tray 310, and the upper support 350, the lower support 360, the bottom surface of the sample tray 310, and the outer ring of the bearing 340 are fixed relative to each other. The upper support 350 and the lower support 360 clamp the bottom surface of the sample tray 310 and fix it to the outer ring of the bearing 340. Simultaneously, the lower support 360 is used to connect to the rotating sample stage 400 so that the rotating sample stage 400 drives the sample tray 310 to rotate. In this embodiment, the stirring component 320 is connected to the connecting component 330, and the connecting component 330 is fixed to the inner ring of the bearing 340. The upper support component 350, the lower support component 360, the sample tray 310, and the outer ring of the bearing 340 are fixed, so that the stirring component 320 remains stationary when the sample tray 310 rotates. In this embodiment, the upper support component 350 includes a first upper support portion 351 and a second upper support portion 352 connected to each other. The outer diameter of the second upper support portion 352 is smaller than the outer diameter of the first upper support portion 351. The upper support component 350 is also provided with an upper receiving hole that passes through the first upper support portion 351 and the second upper support portion 352. The second upper support portion 352 is inserted into the mounting hole. A first gasket 353 is provided between the first upper support portion 351 and the bottom surface of the sample tray 310. In this embodiment, the lower support member 360 includes a first lower support portion 361 and a second lower support portion 362 connected to each other. The outer diameter of the second lower support portion 362 is smaller than the outer diameter of the first lower support portion 361. The lower support member 360 also has a lower receiving hole that passes through the first lower support portion 361 and the second lower support portion 362. The second lower support portion 362 is inserted into the mounting hole. A second gasket 363 is provided between the first lower support portion 361 and the bottom surface of the sample tray 310. The second upper support portion 352 and the second lower support portion 362 are interference-fitted with the mounting hole to achieve relative fixation between the upper support member 350, the lower support member 360 and the sample tray 310. The upper receiving hole includes a first hole section and a second hole section. The inner diameter of the first hole section is larger than the inner diameter of the second hole section, and the first hole section is closer to the lower support member 360 than the second hole section. The bearing 340 is located in the first hole section and abuts against the transition point between the first and second hole sections. The bearing 340 is interference-fitted with the first hole section, achieving relative fixation between the outer ring of the bearing 340 and the upper support member 350. The lower support member 360 abuts against the outer ring of the bearing 340, achieving relative fixation between the bearing 340 and the lower support member 360. Therefore, the bearing 340 is located between the lower support member 360 and the transition point between the first and second hole sections. The first gasket 353 and the second gasket 363 prevent damage to the bottom surface of the sample tray 310.

[0036] Furthermore, the sample carrying assembly 300 also includes a sealing ring 370. The connector 330 has a connecting groove 331, and a connecting shaft 332 is provided in the connecting groove 331. The connecting shaft 332 is relatively fixed to the inner ring of the bearing 340. The upper support 350 is located in the connecting groove 331 and sleeved on the outside of the connecting shaft 332. The sealing ring 370 is located between the groove wall of the connecting groove 331 and the outer side of the upper support 350. In this embodiment, the connecting shaft 332 is inserted into the inner ring of the bearing 340 and has an interference fit with the inner ring of the bearing 340, so that the connecting shaft 332 and the inner ring of the bearing 340 are relatively fixed. By setting the connecting groove 331 in the connector 330, the upper support 350 is located in the connecting groove 331, and the sealing ring 370 is provided between the upper support 350 and the groove wall of the connecting groove 331, a dustproof and waterproof effect is achieved, preventing rapeseed samples, debris, moisture, etc. from entering the instrument interior such as the bearing 340, thus extending the service life. Optionally, the sealing ring 370 is a fluororubber ring. In this embodiment, a sealing groove is provided on the outer side wall of the upper support 350, and the sealing ring 370 is embedded in the sealing groove. In addition, the end face of the connector 330 away from the connecting shaft 332 is an arc surface, which can prevent the sample from accumulating on the top of the connector 330. When the sample is above the connector 330 or above the stirrer 320, it will roll into the sample tray 310 under the action of gravity.

[0037] In addition, the sample carrying assembly 300 also includes a first limiting member 380, a lower support member 360 with a lower receiving hole, a limiting groove at one end of the first limiting member 380 (the limiting groove is non-circular), and the other end of the first limiting member 380 passes through the lower receiving hole and is fixed relative to the inner ring of the bearing 340. A second limiting member 410, which mates with the limiting groove, is provided on the rotating sample stage 400. In use, the sample tray 310 is placed on the rotating sample stage 400, and the second limiting member 410 on the rotating sample stage 400 is inserted into the limiting groove. Because the limiting groove is non-circular, the second limiting member 410 on the rotating sample stage is stationary. Therefore, the first limiting member 380 is stationary when the sample tray 310 rotates. The first limiting member 380 is fixed to the inner ring of the bearing 340, thus keeping the inner ring of the bearing 340 stationary, and consequently, keeping the stirring member 320 stationary. Optionally, the limiting groove is a D-shaped groove.

[0038] The rotating sample stage 400 of this embodiment includes a base 420 and a rotating ring 430. The rotating ring 430 is rotatably connected to the base 420, and the second limiting member 410 is located at the center of the rotating ring 430. When the sample carrying assembly 300 is placed on the rotating sample stage 400, the second limiting member 410 is inserted into the limiting groove of the first limiting member 380. The rotation of the rotating ring 430 drives the sample tray 310 to rotate, while the stirring member 320 remains stationary. The bottom surface of the sample tray 310 in this embodiment is provided with a groove, so that the bottom surface of the sample tray 310 is higher than the bottom surface of its sidewall. The lower support member 360 and the first limiting member 380 are located in the groove. The base 420 is provided with a boss 421, which is located inside the rotating ring 430. The second limiting member 410 is located at the center of the boss 421. The outer diameter of the boss 421 is 1-5 mm smaller than the inner diameter of the sample disk 310. The boss 421 increases the height of the second limiting member 410, preventing the second limiting member 410 from being too long and easily breaking. The boss 421 can extend into the groove of the sample disk 310, serving a positioning function. In addition, the groove wall of the bottom surface of the sample disk 310 in this embodiment is provided with a plurality of circumferentially spaced driven stops 314, and the rotating ring 430 is provided with a plurality of circumferentially spaced driving claws 431, which cooperate with the driven stops 314. When the sample disk 310 is placed on the rotating sample stage 400, the rotating ring 430 is located in the bottom groove of the sample disk 310. Rotating the sample disk 310 causes the driving pawl 431 to abut against the driven stop 314. When the rotating ring 430 rotates, the driving pawl 431 rotates with the rotating ring 430. The driving pawl 431 abuts against the driven stop 314 and thus pushes the sample disk 310 to rotate.

[0039] In this embodiment, the rapeseed near-infrared spectroscopy detection equipment further includes a drive device 700, which is connected to the rotating ring 430 to drive the rotating ring 430 to rotate. Optionally, the drive device 700 includes a motor 710, a drive gear 720, and a driven gear 730. The drive gear 720 is mounted on the output shaft of the motor 710. The driven gear 730 is rotatably connected to the base 420 and connected to the rotating ring 430. The drive gear 720 meshes with the driven gear 730. The rotation of the motor 710 drives the drive gear 720 to rotate, which in turn drives the driven gear 730 to rotate, thereby causing the rotating ring 430 connected to the driven gear 730 to rotate. In this embodiment, the driven gear 730 is located below the base 420, and the rotating ring 430 is located above the base 420. The base 420 has an annular hole, and the driven gear 730 has a through-hole. The through-hole passes through the annular hole and connects to the rotating ring 430. In one specific embodiment, the sample tray 310 includes a base plate 311 and an annular sidewall 312 connected to the edge of the base plate 311. The base plate 311 is a light-transmitting plate. The light source module 100, calibration module 200, and receiving module 500 are located below the rotating sample stage 400, and the sample tray 310 is located above the rotating sample stage 400. In this embodiment, the sample tray 310 is positioned at the top for ease of operation. The bottom surface of the sample tray 310 is made of glass, allowing the detection light emitted by the light source module 100 to pass through the bottom surface of the sample tray 310 and illuminate the sample inside the sample tray 310, so that the signal light generated by the sample under illumination is received by the receiving module 500.

[0040] In this embodiment, the base plate 311 is detachably connected to the annular sidewall 312. The base plate 311 includes a first base plate and a second base plate. The second base plate is provided with a stirring element 320, a connecting element 330, a bearing 340, an upper support element 350, a lower support element 360, and a first limiting element 380. The first base plate is a complete glass plate. By detaching and connecting the base plate 311 to the annular sidewall 312, the base plate 311 can be replaced with either a base plate with a complete glass plate or a base plate 311 with a stirring element 320. When the base plate 311 is a complete glass plate, samples that do not require stirring can be detected. When the base plate 311 has a stirring element 320, it can be used to detect samples that require stirring, thus expanding the detection range of the equipment. In this embodiment, the base plate 311 is higher than the bottom surface of the annular sidewall 312, so that the bottom surface of the sample tray 310 forms a groove to accommodate the lower support element 360 and the first limiting element 380. Furthermore, the driven stop 314 is provided on the inner side of the annular sidewall 312, and the driven stop 314 is located below the base plate 311.

[0041] Optionally, a retaining ring 313 is provided on the inner side of the annular inner wall. The inner diameter of the retaining ring 313 is smaller than the outer diameter of the base plate 311. When the base plate 311 is installed on the annular inner wall, the base plate 311 is located above the retaining ring 313, which facilitates disassembly and assembly. The driven stop block 314 is connected to the retaining ring 313 to improve the load-bearing capacity of the retaining ring 313 and prevent the retaining ring 313 from being crushed after the sample tray 310 is placed in the sample. In this embodiment, the annular sidewall 312, the retaining ring 313, and the driven stop block 314 are integrally manufactured.

[0042] Furthermore, the rapeseed near-infrared spectroscopy detection device of this embodiment also includes an identification module 800. The identification module 800 includes a detection device 810 and an identification device 820. The identification device 820 is disposed on the sample tray 310, and the detection device 810 is used to detect the identification device 820. By setting the identification module 800, different sample carrier components 300 can be identified, thereby identifying the sample type within the sample carrier component 300. This allows the rotating sample stage 400 to use different rotation parameters, and enables the data processing and analysis module to call a model adapted to the sample. It also allows monitoring of the sample carrier component 300, identifying abnormally used sample carrier components 300, ensuring the normal operation of the detection. Optionally, the identification device 820 uses an RFID (Radio Frequency Identification) module.

[0043] like Figure 19 As shown, in this embodiment, yellow rapeseed, blue rapeseed, and black rapeseed are placed in the sample tray 310 from bottom to top. The sample tray 310 is rotated by the rotating sample stage 400 while the stirring component 320 remains stationary. After rotating several times, it can be seen that the yellow rapeseed in the lower layer is brought to the upper layer, and there are yellow rapeseed in both the inner and outer rings, realizing the up-and-down flipping and inside-outside replacement of the sample.

[0044] The other structures in this embodiment are the same as in Embodiment 1, and will not be described again here.

[0045] Example 3 This invention provides a method for detecting rapeseed near-infrared spectroscopy, based on the rapeseed near-infrared spectroscopy detection equipment of Embodiment 1 or Embodiment 2, comprising: The sample to be tested is placed in the sample tray 310. The sample tray 310 is rotated relative to the stirring component 320 by rotating the sample stage 400. During the rotation, the sample tray 310 causes the sample inside to be tumbled and mixed in the axial direction and to generate convection exchange between the center and the edge in the radial direction. During the rotation of the sample disk 310, the detection light emitted from the light source module 100 illuminates the sample on the sample disk 310. The spectrometer 600 collects the light signal generated by the sample after being illuminated by the detection light, which is received by the receiving module 500, according to the preset sampling frequency, and obtains multiple raw spectral data. The calibration module 200 is placed in the optical path of the light source module 100, and wavelength calibration data and background calibration data are collected by the spectrometer 600. The data processing and analysis module converts the original spectral data based on the wavelength calibration data and the background calibration data to obtain converted spectral data. Abnormal spectra are removed from the converted spectral data to obtain valid spectral data. The valid spectral data is preprocessed to obtain preprocessed sample spectral data. Then, based on a preset algorithm, the preprocessed sample spectral data is calculated and the corresponding component content value is output.

[0046] In this embodiment, the sample stage 400 rotates the sample disk 310 relative to the stirring component 320, causing the sample disk 310 to rotate and mix the sample inside it in the axial direction and generate convection exchange between the center and the edge in the radial direction, thereby achieving uniform detection of the entire sample and significantly improving the representativeness and repeatability of the detection results.

[0047] Further, the step of removing abnormal spectra from the converted spectral data to obtain valid spectral data includes: The absorbance value of the original spectral data is detected to be within a preset range. If it exceeds the preset range, the original spectral data is marked as an abnormal spectrum and removed to obtain the first set of filtered spectra. Calculate the correlation coefficient between the original spectral data in the first screened spectral set and the average value of all the original spectral data. If the correlation coefficient is lower than a first preset threshold, mark the original spectral data as an abnormal spectrum and remove it to obtain the second screened spectral set. Calculate the first standard deviation of the original spectral data in the second screening spectral set within the sample characteristic band, and compare it with the second standard deviation of the normal spectrum within the sample characteristic band. If the deviation between the first standard deviation and the second standard deviation exceeds a second preset threshold, the original spectral data is marked as an abnormal spectrum and removed to obtain a third screening spectral set. The original spectral data in the third screening spectral set is valid spectral data.

[0048] During the rotation of the sample disk 310, the spectrometer 600 continuously acquires spectral data at a fixed frequency. For example, approximately 150 raw spectra can be acquired within about 60 seconds of 5 rotations. The acquired spectral data is stored in a data buffer in chronological order. Due to changes in the material state and interference from impurities during the stirring process, a small number of abnormal spectra may be generated. Therefore, in this embodiment, after acquiring multiple raw spectral data through the spectrometer 600, abnormal spectra are first removed from these raw spectral data. This embodiment employs a three-level abnormal detection and removal mechanism: The first level is spectral integrity verification: This checks whether the absorbance value of a single raw spectral data point is within a preset range. Spectra outside this range are directly marked as abnormal and removed. The remaining raw spectral data form the first set of selected spectra. In this embodiment, the preset range for absorbance values ​​is -0.5 to 3.0.

[0049] The second level is the spectral correlation test: The correlation coefficient between each original spectral data in the first screened spectral set and the average of all original spectral data is calculated. Spectra with correlation coefficients lower than a first preset threshold are marked as abnormal spectra and removed. The remaining original spectral data form the second screened spectral set. This step can identify abnormal spectra caused by abrupt changes in the material's stacking state. In this embodiment, the first preset threshold is 0.95.

[0050] The third level is the spectral dispersion test: The standard deviation of each original spectral data in the characteristic band of the second screening spectral set is calculated and compared with the standard deviation of the normal spectrum. Spectra with a deviation exceeding a second preset threshold are marked as abnormal and removed. The remaining original spectral data form the third screening spectral set. In this embodiment, the second preset threshold is that the deviation of the standard deviation of each original spectral data in the characteristic band from the standard deviation of the normal spectrum is three times the standard deviation of the normal spectrum. Here, "normal spectrum" refers to the "standard appearance" statistically derived from a large number of qualified samples, and its standard deviation from the sample's characteristic band is a measure of the "distance between the sample and the standard." Normal spectrum can be obtained by calculating the average spectrum or extracting typical characteristic spectra from normal samples. Calculating the average spectrum involves collecting a batch (e.g., more than 30) of "normal" samples verified by standard methods, collecting their spectra under the same conditions, and averaging the intensity at each wavelength point to obtain an average spectrum. This spectral curve represents the "standard appearance" of this type of qualified sample. The standard deviation of the normal spectrum in the characteristic band of a sample is an objective benchmark used to measure "how much difference is considered normal," that is, "the inherent fluctuation range of the spectrum of a qualified product." First, the spectral data of the qualified samples for establishing the "normal spectrum" are obtained. For a selected "characteristic band" (e.g., the characteristic absorption peak of water at 1450nm), the standard deviation of the spectral intensity of all normal samples at each wavelength point within this band is calculated. If necessary, an average of the standard deviations within this band is taken as an overall fluctuation index for that characteristic band. This yields the standard deviation of the normal spectrum in the sample's characteristic band, representing the inherent, normal fluctuation of a normal sample in this characteristic band due to minor natural differences in composition, particle size, etc. In this embodiment, if the standard deviation of the original spectral data in the characteristic band falls within the normal fluctuation range (e.g., within ±3 standard deviations), it can be judged as "normal." If it exceeds the normal fluctuation range, it can be judged as "abnormal."

[0051] The effective spectral data is preprocessed to obtain preprocessed sample spectral data, including eliminating physical interference and highlighting chemical information. Specifically, this includes: derivative methods (first derivative, second derivative), smoothing methods (Savitzky-Golay smoothing, moving average smoothing), baseline correction methods, and normalization methods (SNV, MSC). Multivariate scattering correction (MSC) or standard normal variable transformation (SNV) is used to eliminate the influence of particle scattering; derivative methods are used to separate overlapping peaks and correct baseline drift; and smoothing methods are used to improve the signal-to-noise ratio. These methods all fundamentally change the shape and value of the original spectrum.

[0052] The process of preprocessing the effective spectral data to obtain preprocessed sample spectral data, and then calculating and outputting the corresponding component content values ​​based on a preset algorithm, includes: The effective spectral data are preprocessed to obtain preprocessed sample spectral data; The preprocessed sample spectral data are fused to obtain fused spectral data; The fused spectral data is calculated based on a preset algorithm, and the corresponding component content values ​​are output.

[0053] The preprocessed sample spectral data are fused to obtain fused spectral data, which is processed using one of the following three fusion methods: The first method is the arithmetic mean method: the arithmetic mean of all effective spectra is calculated at each wavelength to obtain a representative spectrum. This method is simple and efficient, and suitable for samples with good homogeneity.

[0054] The second method is the weighted average method: the weight is calculated based on the quality score of each spectrum (based on indicators such as signal-to-noise ratio and spectral consistency), and the spectrum with the higher weight contributes more to the final result.

[0055] The third method is the median fusion method: the median of all valid spectra is taken at each wavelength as the fusion result. This method is more robust to outliers.

[0056] This embodiment uses the arithmetic mean method to process the effective spectrum. In actual testing, this method can effectively balance detection efficiency and result accuracy. The representative spectrum after fusion processing is then subjected to further preprocessing and model prediction to obtain the final detection result. The sample in this embodiment is rapeseed. The spectrum obtained after fusing the effective spectral data of rapeseed is shown in Figure 20.

[0057] In calculating and outputting the corresponding component content values ​​based on the fused spectral data using a preset algorithm, the establishment of the preset algorithm includes: Spectral range optimization: By traversing different spectral intervals, a model encompassing all wavelengths and characteristic band selections is established. After comprehensively comparing the model performance, the optimal spectral range is determined. This allows for the selection of the spectral interval with the strongest correlation to the measured component and the least noise.

[0058] Principal Factor Count Optimization: The cross-validation method is used to systematically optimize the number of principal factors in the partial least squares (PLS) regression model, achieving the best balance between model complexity and overfitting. This prevents the model from underfitting due to too few principal factors, while also preventing the model from overfitting due to too many principal factors, which might lead to the inclusion of noise as a regularity, resulting in poorer predictions for new samples.

[0059] Joint optimization of preprocessing and modeling: The selection of preprocessing algorithms and the optimization of modeling parameters are jointly optimized, and the globally optimal combination of parameters is found through methods such as grid search or genetic algorithms.

[0060] Model performance validation: The optimized model is rigorously validated using an independent test set to evaluate its predictive power, robustness, and generalization performance.

[0061] Through the aforementioned process of establishing the preset algorithm, an optimal partial least squares (PLS) quantitative analysis model can be obtained. Inputting the fused spectral data into this quantitative analysis model will output the component content values. The performance of the prediction model in this embodiment is shown in Table 1.

[0062] Table 1. Detection performance of rapeseed component content

[0063] According to Table 1, the preset algorithm model in this embodiment exhibits excellent performance and accurate predictions, achieving high precision in the detection of constant components. Therefore, this embodiment demonstrates outstanding quantitative detection results for multiple key quality indicators of rapeseed, meeting the needs for rapid and non-destructive analysis in various scenarios.

[0064] In addition, in this embodiment, 10 rapeseed samples with different component contents were selected, and each sample was measured 5 times. The range of each component in the 5 measurements was taken as the repeatability error. The comparison between the stirring effect of the present invention and the effect without stirring is shown in Tables 2 and 3. It can be seen that the measurement repeatability of the sample was greatly improved after stirring.

[0065] Table 2 Repeatability error of each sample after stirring

[0066] Table 3. Repeatability error of each sample without stirring

[0067] Based on the test results in Tables 2 and 3, it is evident that stirring significantly improves the accuracy and stability of near-infrared spectroscopy for rapeseed. Direct detection without stirring leads to significant repeatability deviations in the predicted values ​​of key indicators such as oil content and erucic acid. For example, the range of oil content measurements for sample No. 10 without stirring was as high as 1.1%, far exceeding the range of 0.4% after stirring. Stirring resulted in better consistency across all samples, with a significant reduction in fluctuations in the test data.

[0068] In summary, this invention provides a near-infrared spectroscopy detection device and method for rapeseed. The sample carrying component 300 includes a sample tray 310 and a stirring element 320. The relative rotation of the sample tray 310 and the stirring element 320 allows the sample in the sample tray 310 to be turned over, achieving the exchange of upper and lower layers and uniform mixing of inner and outer rings. This redistributes the seeds at different layers and positions within the sample cup without altering the overall sample loading state, enabling the detection beam to quickly and continuously scan a wider sample area. This effectively balances the requirements of information representativeness, operational efficiency, and detection repeatability, ensuring reliable test results and stable experimental repeatability, and reflecting the true data of particulate matter to a greater extent. Furthermore, this invention uses a rotating sample stage 400 to rotate the sample tray 310 while the stirring element 320 remains stationary. The rotation axis of the sample tray 310 is offset from the optical axis of the light source module 100, preventing the spectrometer 600 from acquiring signals from the stirring element 320 and affecting data integrity, thus improving the accuracy and reliability of the detection. Furthermore, the central axis of the receiving module 500 of this invention forms a certain angle with the optical axis of the light source module 100, effectively avoiding interference from specular reflection light and achieving better overlap between the irradiated light spot and the received light spot, ensuring a sufficiently large light spot area. In addition, this invention employs a three-level anomaly detection mechanism—spectral integrity, correlation, and dispersion—to eliminate abnormal spectra from multiple acquired spectral data. This effectively eliminates outlier data caused by sample state fluctuations, sample loading differences, or accidental interference, improving the consistency and reliability of spectral data, thereby enhancing the accuracy and robustness of subsequent quantitative or qualitative analysis models.

[0069] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention.

Claims

1. A near-infrared spectroscopy detection device for rapeseed, characterized in that, include: A light source module (100) is used to provide detection light to illuminate the sample; Calibration module (200) for wavelength calibration and background calibration; The sample carrier assembly (300) includes a sample tray (310) and a stirrer (320), wherein the sample tray (310) is used to hold the sample and the stirrer (320) is disposed inside the sample tray (310); The rotating sample stage (400) is used to drive the sample disk (310) to rotate relative to the stirring component (320), so that the sample disk (310) drives the sample inside to tumble and mix in the axial direction and generate convection exchange between the center and the edge in the radial direction during the rotation process, and the rotation center axis of the sample disk (310) is offset from the optical axis of the light source module (100). The receiving module (500) is used to receive the light signal generated by the sample after it is irradiated by the detection light; A spectrometer (600) is used to receive the optical signal transmitted by the receiving module (500) and generate corresponding spectral data; A data processing and analysis module is used to process and analyze the spectral data.

2. The rapeseed near-infrared spectroscopy detection device according to claim 1, characterized in that, The top surface of the stirring element (320) is an inclined surface, and the bottom surface of the stirring element (320) is provided with a through groove (324). The through groove (324) extends inclinedly from one end of the stirring element (320) away from the rotation center axis of the sample plate (310) to the end closer to the rotation center axis of the sample plate (310), or the through groove (324) extends inclinedly from one end of the stirring element (320) close to the rotation center axis of the sample plate (310) to the end away from the rotation center axis of the sample plate (310).

3. The rapeseed near-infrared spectroscopy detection device according to claim 1, characterized in that, The sample carrying assembly (300) further includes a connector (330), a bearing (340), an upper support (350), and a lower support (360). The sample tray (310) has a mounting hole at its center. The bearing (340) is located in the mounting hole. The stirring component (320) is connected to the connector (330). The connector (330) is fixed relative to the inner ring of the bearing (340). The upper support (350) and the lower support (360) are respectively located on both sides of the bottom surface of the sample tray (310). The upper support (350), the lower support (360), the bottom surface of the sample tray (310), and the outer ring of the bearing (340) are fixed relative to each other.

4. The rapeseed near-infrared spectroscopy detection device according to claim 3, characterized in that, The sample carrier assembly (300) also includes a sealing ring (370), the connector (330) is provided with a connecting groove (331), a connecting shaft (332) is provided in the connecting groove (331), the connecting shaft (332) is fixed relative to the inner ring of the bearing (340), the upper support (350) is located in the connecting groove (331) and sleeved on the outside of the connecting shaft (332), and the sealing ring (370) is located between the groove wall of the connecting groove (331) and the outer side of the upper support (350).

5. The rapeseed near-infrared spectroscopy detection device according to claim 3 or 4, characterized in that, The sample carrier assembly (300) further includes a first limiting member (380), the lower support member (360) is provided with a lower receiving hole, one end of the first limiting member (380) is provided with a limiting groove, the limiting groove is non-circular, the other end of the first limiting member (380) passes through the lower receiving hole and is fixed relative to the inner ring of the bearing (340), and the rotating sample stage (400) is provided with a second limiting member that cooperates with the limiting groove.

6. The rapeseed near-infrared spectroscopy detection device according to claim 1, characterized in that, The sample tray (310) includes a base plate (311) and an annular sidewall (312) connected to the edge of the base plate (311). The base plate (311) is a light-transmitting plate. The light source module (100), the calibration module (200), and the receiving module (500) are located below the rotating sample stage (400), and the sample tray (310) is located above the rotating sample stage (400).

7. The rapeseed near-infrared spectroscopy detection device according to claim 1, characterized in that, The central axis of the receiving module (500) is set at a certain angle to the optical axis of the light source module (100).

8. The rapeseed near-infrared spectroscopy detection device according to claim 1, characterized in that, It also includes an identification module (800), which includes a detection device (810) and an identification device (820). The identification device (820) is disposed on the sample tray (310), and the detection device (810) is used to detect the identification device (820).

9. A method for detecting rapeseed near-infrared spectroscopy, based on the rapeseed near-infrared spectroscopy detection equipment according to any one of claims 1-8, characterized in that, include: The sample to be tested is placed in the sample tray (310). The sample tray (310) is rotated relative to the stirring piece (320) by rotating the sample stage (400). During the rotation, the sample tray (310) causes the sample inside to be tumbled and mixed in the axial direction and to generate convection exchange between the center and the edge in the radial direction. During the rotation of the sample disk (310), the detection light emitted by the light source module (100) illuminates the sample on the sample disk (310), and the spectrometer (600) collects the light signal generated by the sample after being illuminated by the detection light and received by the receiving module (500) at a preset sampling frequency, thereby obtaining multiple original spectral data. The calibration module (200) is placed in the optical path of the light source module (100), and wavelength calibration data and background calibration data are collected by the spectrometer (600); The data processing and analysis module converts the original spectral data based on the wavelength calibration data and the background calibration data to obtain converted spectral data. Abnormal spectra are removed from the converted spectral data to obtain valid spectral data. The valid spectral data is preprocessed to obtain preprocessed sample spectral data. Then, based on a preset algorithm, the preprocessed sample spectral data is calculated and the corresponding component content value is output.

10. The method for detecting rapeseed using near-infrared spectroscopy according to claim 9, characterized in that, The step of removing abnormal spectra from the converted spectral data to obtain valid spectral data includes: The absorbance value of the original spectral data is detected to be within a preset range. If it exceeds the preset range, the original spectral data is marked as an abnormal spectrum and removed to obtain the first set of filtered spectra. Calculate the correlation coefficient between the original spectral data in the first screened spectral set and the average value of all the original spectral data. If the correlation coefficient is lower than a first preset threshold, mark the original spectral data as an abnormal spectrum and remove it to obtain the second screened spectral set. Calculate the first standard deviation of the original spectral data in the second screening spectral set within the sample characteristic band, and compare it with the second standard deviation of the normal spectrum within the sample characteristic band. If the deviation between the first standard deviation and the second standard deviation exceeds a second preset threshold, the original spectral data is marked as an abnormal spectrum and removed to obtain a third screening spectral set. The original spectral data in the third screening spectral set is valid spectral data.

Images