Spectrum detection system and method

Abstract

The invention provides a spectrum detection system and method, and the system comprises a self-adaptive clamping and conveying device which comprises two conveying units which are arranged in a V-shaped included angle space manner in the conveying direction of the self-adaptive clamping and conveying device, and a vertically-through light path penetrating groove is formed between the bottoms of the two conveying units; the dynamic adjustable transmission optical module comprises a light source assembly and a detection assembly, the light source assembly and the detection assembly are arranged on the upper side and the lower side of the light path penetrating groove respectively, and the light source assembly is used for emitting detection light penetrating through the light path penetrating groove and a detected object. The detection assembly is used for receiving the detection light and collecting a spectral signal penetrating through a detected object; and the non-contact laser ranging diameter measuring module comprises a laser ranging unit, and the laser ranging unit is arranged towards the light path penetrating groove. The problems that in an existing transmission detection system, a clamp is difficult to stably clamp and convey products, and the optical path length cannot be dynamically, stably and controllably achieved are solved.

Description

Technical Field

[0001] This invention relates to the field of transmission detection, and more particularly to a spectral detection system and method. Background Technology

[0002] Visible and near-infrared (Vis / NIR) transmission spectroscopy is a powerful tool for assessing the internal quality of products. For example, in evaluating agricultural products, internal quality includes factors such as sugar content, dry matter, maturity, uniformity of composition, internal defects, and pests and diseases. This technique follows the Beer-Lambert Law, which states that absorbance is directly proportional to the concentration of the substance and the optical path length. The optical path length includes the actual thickness or diameter of the object through which the light beam penetrates.

[0003] In a laboratory setting, the optical path length can be precisely controlled by slicing, preparing, or using standard cuvettes, thus achieving high-precision quantitative analysis. However, in high-speed, online detection scenarios, the samples being tested often have irregular shapes and significant size variations. Regular clamps in transmission detection systems struggle to stably hold and transport the products, rigid transport devices are prone to damaging the sample surface, and they cannot adapt to differences in sample diameter. This makes the optical path length a huge, highly random, and unpredictable variable that decisively affects the measurement results.

[0004] To address the optical path length issue, some solutions attempt to integrate independent online thickness measurement modules (such as optical interferometry and structured light 3D scanning). These solutions are typically complex and costly, requiring additional independent light sources, detectors, sophisticated scanning mechanisms, and high-performance image processing or 3D point cloud algorithms. This significantly increases the hardware cost and software development difficulty of the entire system, making its maintenance more complex. Summary of the Invention

[0005] This invention provides a spectral detection system and method that solves the problems of unstable clamping and conveying of products and the inability to achieve dynamic stability and controllability of optical path length in existing transmission detection systems.

[0006] This invention provides a spectral detection system, comprising:

[0007] An adaptive clamping and conveying device includes two conveying units arranged in a V-shaped angle along their conveying direction, and a light path penetration groove that is open at both the top and bottom is provided between the bottoms of the two conveying units.

[0008] A dynamically adjustable transmission optical module includes a light source assembly and a detection assembly. The light source assembly and the detection assembly are respectively arranged on the upper and lower sides of the optical path penetration slot. The light source assembly is used to emit detection light rays that pass through the optical path penetration slot and the object under test. The detection assembly is used to receive the detection light rays and collect the spectral signal that penetrates the object under test.

[0009] A non-contact laser ranging diameter measurement module includes a laser ranging unit, which is positioned facing the optical path penetration groove;

[0010] The data processing unit is electrically connected to the dynamically adjustable transmission optical module and the non-contact laser ranging diameter measurement module, respectively.

[0011] The control unit is connected to the adaptive clamping and conveying device, the dynamically adjustable transmission optical module, and the non-contact laser ranging and diameter measurement module, respectively.

[0012] In one embodiment of the present invention, the data processing unit is used to acquire the spectral signal and the ranging data collected by the non-contact laser ranging diameter measurement module, process the spectral signal and the ranging data to calculate the thickness of the object under test, correct the spectral signal based on the thickness or predict the internal component content of the object under test, and generate a defect judgment result of the object under test.

[0013] In one embodiment of the present invention, the data processing unit includes a synchronous data processing unit and a data fusion processing unit;

[0014] The synchronous data processing unit is used to calculate the instantaneous thickness D(t) = H0 - H1(t) - Δh of the measured object based on the ranging data, where H0 is the vertical distance from the laser ranging unit transmitter to the horizontal plane where the optical path penetration slot is located, which is the laser ranging reference value; H1(t) is the real-time vertical distance from the laser ranging unit transmitter to the measured object; and Δh is the deformation of the elastic support component after being compressed. The ranging data includes the laser ranging reference value and the real-time vertical distance.

[0015] The data fusion processing unit is used to correct the absorbance spectral value calculated from the spectral signal by using the instantaneous thickness of the object under test as an optical path length parameter based on the Beer-Lambert law, or to substitute the instantaneous thickness of the object under test into the internal quality quantitative prediction model to calculate the internal component content parameter of the object under test.

[0016] In one embodiment of the present invention, the adaptive clamping and conveying device further includes a conveyor belt support, the conveying unit is a conveyor belt, the conveyor belt is slidably arranged in a V-shape on the conveyor belt support, and the optical path penetration groove is opened at the bottom of the V-shape of the conveyor belt support.

[0017] In one embodiment of the present invention, an adjustment structure is rotatably provided on the top of the conveyor belt support. The adjustment structure is located at the entrance of the adaptive clamping conveyor device and is offset from the optical path penetration groove. A gap is provided between the bottom of the adjustment structure and the conveyor belt.

[0018] In one embodiment of the present invention, a pressure sensing reference mechanism is further included. The pressure sensing reference mechanism includes an elastic support component and a pressure sensor. The two sides of the conveyor belt support away from the optical path penetration groove are mounted on the frame through the elastic support component. The pressure sensor is disposed between the conveyor belt and the conveyor belt support.

[0019] In one embodiment of the present invention, the elastic support assembly includes a lifting device and a constant force spring, the constant force spring being disposed on the top of the lifting device, and the lifting device being controlled and connected to the control unit.

[0020] In one embodiment of the present invention, an installation beam is provided above the adaptive clamping and conveying device, and a mounting frame is provided on the installation beam. The laser ranging unit and the detection component are both mounted on the mounting frame.

[0021] The mounting beam has a first adjustment groove extending in the left-right direction, and the mounting frame has a second adjustment groove extending in the up-down direction. The first adjustment groove and the second adjustment groove are connected by fastening bolts.

[0022] In one embodiment of the present invention, the detection component is provided with sample sensors at intervals on the side facing the feed inlet of the adaptive clamping conveyor, and the sample sensors are arranged facing the optical path penetration groove.

[0023] In another aspect, the present invention provides a spectral detection method, comprising:

[0024] An adaptive clamping and conveying device is used to transport the object under test so that the object under test passes through the detection area;

[0025] The spectral signal of the object under test is acquired by a dynamically adjustable transmission optical module.

[0026] The distance measurement data of the object being measured is acquired by a non-contact laser ranging diameter measurement module;

[0027] The data processing unit acquires the spectral signal and the ranging data, calculates the thickness of the object under test based on the spectral signal and the ranging data, corrects the spectral signal or predicts the internal component content of the object under test based on the thickness, and generates a defect judgment result for the object under test.

[0028] In one embodiment of the present invention, before the step of using the adaptive clamping and conveying device to convey the object under test, the method further includes:

[0029] Offline calibration is performed, and the vertical distance from the laser ranging unit's emitting end to the horizontal plane where the optical path penetration groove is located is obtained through the non-contact laser ranging diameter measurement module, which serves as the laser ranging reference value.

[0030] The corresponding spectral signal is acquired by the dynamically adjustable transmission optical module as a reference light intensity.

[0031] In one embodiment of the present invention, the step of using an adaptive clamping and conveying device to convey the object under test includes:

[0032] Place the object to be tested into the inlet of the adaptive clamping and conveying device.

[0033] The object under test is pushed to the V-shaped angle line between the two transmission units and the adjustment structure through two transmission units arranged in a V-shape, thereby automatically centering and correcting the posture of the object under test.

[0034] In one embodiment of the present invention, the thickness is an instantaneous thickness D(t), which is calculated by the formula: D(t) = H0 - H1(t) - Δh, where H0 is the laser ranging reference value, H1(t) is the real-time vertical distance collected by the non-contact laser ranging diameter measurement module, and Δh is the deformation of the elastic support component in the adaptive clamping and conveying device after being compressed. The ranging data includes the laser ranging reference value and the real-time vertical distance.

[0035] In one embodiment of the present invention, correcting the spectral signal based on the thickness includes:

[0036] Calculate the absorbance spectral value A(λ) = -log10(I(λ) / I0(λ)), where I(λ) is the real-time spectral signal transmitted through the object being tested, acquired by the detection component, and I0(λ) is the reference light intensity;

[0037] The instantaneous thickness of the object under test is used as the optical path length parameter to correct the absorbance spectral value. The correction formula is: A _corrected =k*A(λ) / D(t), where A _corrected The absorbance spectral value is the corrected value, and k is a proportionality constant or model coefficient.

[0038] In one embodiment of the present invention, predicting the internal component content of the object being tested includes:

[0039] Substitute the spectral signal and the thickness into the internal quality quantitative prediction model to output the internal component content of the measured object;

[0040] The internal quality quantitative prediction model is trained through the following steps:

[0041] Prepare several test object samples covering different sizes and internal qualities;

[0042] The spectral signals and ranging data of each sample are acquired by the dynamically adjustable transmission optical module and the non-contact laser ranging diameter measurement module to obtain spectral-thickness data.

[0043] Determine the true values ​​of the internal component content of each sample;

[0044] Using the spectral-thickness data as input variables and the true values ​​of the internal component content as output variables, the internal quality quantitative prediction model is trained using partial least squares regression, support vector machine algorithm, or deep learning algorithm.

[0045] In one embodiment of the present invention, generating the defect judgment result of the tested object includes:

[0046] Substitute the spectral signal and the thickness into the internal defect discrimination model to output the defect judgment result of the tested object;

[0047] The internal defect discrimination model is trained through the following steps:

[0048] Prepare several test object samples covering different sizes and internal qualities;

[0049] The spectral signals and ranging data of each sample are acquired by the dynamically adjustable transmission optical module and the non-contact laser ranging diameter measurement module to obtain spectral-thickness data.

[0050] Manually determine the type of internal defect in each sample;

[0051] Using the spectral-thickness data as input variables and the internal defect category as output variables, the internal defect discrimination model is trained using deep learning, logistic regression, or SVM classifiers.

[0052] As can be seen from the above solutions, the advantages of the present invention are:

[0053] The spectral detection system provided by this invention utilizes a V-shaped adaptive clamping and conveying device and a conveyor belt unit with a V-shaped spatial layout. The bottom features a vertically and horizontally permeable optical path penetration groove. Combined with an adjustment structure at the entrance, it automatically pushes the object being measured to the center of the V-shaped angle upon entry, ensuring consistent detection position. Through the spatially coordinated layout of a dynamically adjustable transmission optical module and a non-contact laser ranging module, the light source and detection components in the dynamically adjustable transmission optical module are located on the upper and lower sides of the optical path penetration groove, respectively, emitting and receiving spectral signals penetrating the object being measured. The laser ranging unit, positioned adjacent to the optical path penetration groove, simultaneously performs non-contact ranging on the upper surface of the object. Combined with the lower surface reference displacement obtained by the pressure sensing reference mechanism, spatial synchronization of "spectral characteristic detection" and "diameter measurement" is achieved, overcoming the limitations of traditional single-dimensional detection. Attached Figure Description

[0054] Figure 1 This is a three-dimensional structural diagram of the spectral detection system of Embodiment 1 of the present invention;

[0055] Figure 2 for Figure 1 A schematic diagram of the planar structure from the left-viewing angle;

[0056] Figure 3 for Figure 1 A schematic diagram of the planar structure from the rear view angle;

[0057] Figure 4 for Figure 1 Internal cross-sectional view of the spectral detection chamber;

[0058] Figure 5 for Figure 4 A schematic diagram showing the layout and orientation of the light source and detection components for the rear-view angle.

[0059] Figure 6 This is a schematic flowchart of the spectral detection method provided in Embodiment 2 of the present invention.

[0060] The attached figures are labeled as follows:

[0061] 1: Adaptive clamping and conveying device;

[0062] 11: Transmission unit;

[0063] 12: Conveyor belt support;

[0064] 13: Optical path penetration groove;

[0065] 14: Drive unit;

[0066] 15: Adjust the structure;

[0067] 151: Adjust the roller;

[0068] 152: Shaft;

[0069] 153: Bearing housing;

[0070] 21: Light source assembly;

[0071] 22: Detection components;

[0072] 23: Sample sensor;

[0073] 31: Laser ranging unit;

[0074] 311: Laser rangefinder sensor;

[0075] 4: Rack;

[0076] 41: Adjustable outriggers;

[0077] 42: Adjusting screw;

[0078] 5: Spectral detection chamber;

[0079] 51: Maintenance cover;

[0080] 52: Electrically controlled push rod;

[0081] 6: Install the crossbeam;

[0082] 61: First mounting hole;

[0083] 62: First adjusting groove;

[0084] 7: Mounting bracket;

[0085] 71: Second adjustment groove. Detailed Implementation

[0086] It should be noted that, in this application, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, 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.

[0087] In the absence of further restrictions, an element defined by the phrase "comprising a..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0088] Example 1

[0089] refer to Figures 1 to 3 As shown, one embodiment of the present invention provides a spectral detection system, wherein... Figure 1 A three-dimensional structural schematic diagram of the spectral detection system is shown; Figure 2 for Figure 1 A schematic diagram of the planar structure from the left-viewing angle; Figure 3 for Figure 1 A schematic diagram of the planar structure from the rear view angle.

[0090] A spectral detection system includes an adaptive clamping and conveying device 1, a dynamically adjustable transmission optical module, a non-contact laser ranging diameter measurement module, a data processing unit, a control unit, a frame 4, and a spectral detection chamber 5.

[0091] The adaptive clamping and conveying device 1 includes two conveying units 11 arranged in a V-shape along its conveying direction, with a vertically permeable optical path penetration groove 13 between the bottoms of the two conveying units 11. The adaptive clamping and conveying device 1 and the spectral detection chamber 5 are mounted on the frame 4, and the adaptive clamping and conveying device 1 passes through the spectral detection chamber 5.

[0092] The dynamically adjustable transmission optical module includes a light source assembly 21 and a detection assembly 22. The light source assembly 21 and the detection assembly 22 are respectively arranged on the upper and lower sides of the optical path penetration groove 13. The light source assembly 21 is used to emit detection light that passes through the optical path penetration groove and the object under test. The detection assembly 22 is used to receive the detection light and collect the spectral signal that penetrates the object under test.

[0093] The non-contact laser ranging diameter measurement module includes a laser ranging unit 31, which is disposed facing the optical path penetration slot 13. In one embodiment, the laser ranging unit 31 includes two laser ranging sensors 311 with different wavelengths.

[0094] The control unit is connected to the adaptive clamping and conveying device, the dynamically adjustable transmission optics module, and the non-contact laser ranging and diameter measurement module, respectively, to control these devices to perform corresponding actions. In one embodiment, the control unit is a PLC controller.

[0095] The data processing unit is electrically connected to the dynamically adjustable transmission optical module and the non-contact laser ranging diameter measurement module, respectively, and is communicatively connected to the control unit.

[0096] The adaptive clamping and conveying device 1 includes two conveying units 11 arranged in a V-shaped spatial layout along its conveying direction. A light path penetration groove 13 that is open at the bottom of the two conveying units 11 is provided. The adaptive clamping and conveying device is used to convey the object to be measured and automatically clamp and center the object to be measured.

[0097] In one embodiment, specifically, such as Figure 4 As shown, Figure 4 for Figure 1 The image shows an internal cross-sectional view of the spectral detection chamber 5. The adaptive clamping and conveying device 1 also includes a conveyor belt support 12, which has a V-shaped structure and is mounted on the frame 4, passing through the spectral detection chamber 5. The conveying unit 11 is a conveyor belt, which is slidably arranged in a V-shape on the conveyor belt support 12. The optical path penetration groove 13 is formed at the bottom of the V-shape of the conveyor belt support 12. The conveyor belt 11 is driven by a driving device 14 to rotate around the conveyor belt support 12. In one embodiment, the driving device 14 is a geared motor.

[0098] In order to achieve automatic centering of the object under test on the adaptive clamping conveyor, in this embodiment, the conveyor belt support 12 rotates and has an adjustment structure 15 across its top, which is located at the entrance of the adaptive clamping conveyor and is offset from the light path penetration groove 13. There is a gap between the bottom of the adjustment structure 15 and the conveyor belt 11 for the object under test to pass through. The adjustment roller is used to push the object under test on the conveyor belts 11 on both sides to the V-shaped angle line between the two conveyor belts 11.

[0099] Further reference Figure 2 As shown, in one embodiment, the adjustment structure includes an adjustment roller 151, which is fixedly mounted on the conveyor belt support 12 via a shaft 152 and a bearing seat 153.

[0100] In one embodiment, the spectral detection system further includes a pressure sensing reference mechanism, which comprises an elastic support assembly (not shown) and a pressure sensor (not shown). The two sides of the conveyor belt support 12 away from the optical path penetration groove 13 are mounted on the frame 4 via the elastic support assembly. The pressure sensor is disposed between the conveyor belt 11 and the conveyor belt support 12. In one embodiment, the elastic support assembly includes a lifting device and a constant force spring. The constant force spring is disposed on top of the lifting device, and the lifting device is controlled and connected to the control unit. The lifting device can be a cylinder, hydraulic cylinder, etc.

[0101] In one embodiment, further reference is made to Figure 4As shown, in the spectral detection chamber 5, an installation beam 6 is provided above the adaptive clamping and conveying device 1, and an installation frame 7 is provided on the installation beam 6. The laser ranging unit 31 and the detection component 22 are both installed on the installation frame 7.

[0102] The mounting beam 6 is an L-shaped angle steel structure. A first mounting hole 61, open at both ends, is located in the middle of the mounting beam 6. First adjustment slots 62, extending horizontally and open at both ends, are located on both sides of the mounting beam 6. Mounting brackets 7 are mounted on both the first mounting hole 61 and the first adjustment slots 62. Second adjustment slots 71, extending vertically and open at both ends, are located on the mounting brackets 7. The first adjustment slots 62, the second adjustment slots 71 on both sides, the first mounting hole 61, and the second adjustment slot 71 in the middle are all connected by fastening bolts, allowing adjustment of the vertical position of the central mounting bracket 7 and the horizontal and vertical positions of the mounting brackets 7 on both sides. The right-side mounting bracket 7 is designed for easy observation of the structure of the mounting beam 6. Figure 4 It has been hidden.

[0103] In one embodiment, the dynamically adjustable transmission optics module includes a light source assembly 21 and a detection assembly 22 respectively disposed on the upper and lower sides of the optical path penetration slot 13. The light source assembly 21 emits detection light that passes through the optical path penetration slot 13 and the object under test. The detection assembly 22 receives the detection light and collects the spectral signal penetrating the object under test. The light source assembly 21 can be a monochromatic or multi-band light source with a specific wavelength range (such as an LED or laser diode), or a broadband light source (such as a halogen lamp or NIR lamp). The detection assembly 22 receives the detection light and collects the spectral signal penetrating the object under test. The detection assembly 22 can be a single-point detector (such as a photodiode or photomultiplier tube) or an array detector (such as a CCD or CMOS).

[0104] In one embodiment, further reference is made to Figure 4 , Figure 5 As shown, Figure 5 for Figure 4 A schematic diagram showing the layout and orientation of the light source assembly and detection assembly from a rear-view angle. Sample sensors 23 are spaced apart on the side of the detection assembly 22 facing the feed inlet of the adaptive clamping conveyor, and these sample sensors 23 are positioned towards the light path penetration groove 13. The sample sensors 23 are used to sense whether a measured object is passing below them; the sample sensors 23 can be proximity switches.

[0105] Further reference Figure 4 , Figure 5As shown, in one embodiment, the non-contact laser ranging diameter measurement module includes a laser ranging unit 31, which is disposed on the side of the sample sensor 23 near the detection component 22. The left and right mounting brackets 7 are symmetrically arranged about the optical path penetration groove 16. The sample sensor 23, the laser ranging unit 31, and the detection component 22 are all sequentially mounted on the mounting brackets 7 along the direction of movement of the object being measured, and all three components face the optical path penetration groove 13.

[0106] In one embodiment, the laser ranging unit 31 includes two laser ranging sensors with different wavelengths. By using the reflected signals of the two wavelengths, the measurement stability under complex surface conditions is improved. The laser ranging unit 31 is mounted on a liftable mounting frame 7. A quick-release structure between the mounting frame 7 and the mounting beam 6 allows for switching between a 30-300 mm measurement range to accommodate objects of different sizes. This laser ranging unit 31 is used for non-contact detection of the upper surface of the object during transport, measuring the absolute height of the object relative to a reference benchmark near the optical path penetration point in real time and synchronously. A composite measurement architecture of "laser non-contact ranging + pressure sensing benchmark calibration" is adopted to achieve real-time and accurate detection of the diameter of the object during transport. Its core design logic is as follows: the laser captures the relative distance between feature points on the upper surface of the object and a reference plane (e.g., the horizontal plane where the optical path penetration groove is located is the reference plane). Combined with pressure sensing, the reference displacement of the lower surface of the object due to gravity is obtained. The difference between the two is directly output as the diameter value after geometric model conversion.

[0107] In one embodiment, the data processing unit is electrically connected to both the dynamically adjustable transmission optics module and the non-contact laser ranging diameter measurement module, and is communicatively connected to the control unit. The data processing unit acquires the spectral signal and the ranging data collected by the non-contact laser ranging diameter measurement module, processes the spectral signal and ranging data to calculate the thickness of the object under test, corrects the spectral signal based on the thickness or predicts the internal composition content of the object under test, and generates a defect judgment result for the object under test.

[0108] Specifically, in one embodiment, the data processing unit includes a synchronous data processing unit and a data fusion processing unit. The synchronous data processing unit calculates the instantaneous thickness D(t) = H0 - H1(t) - Δh of the measured object based on the ranging data, where H0 is the vertical distance from the laser ranging unit's transmitter to the horizontal plane where the optical path penetration groove is located, serving as the laser ranging reference value, and the horizontal plane where the optical path penetration groove is located is selected as the reference plane; H1(t) is the real-time vertical distance from the laser ranging unit's transmitter to the measured object; and Δh is the deformation of the elastic support component after being compressed. The ranging data includes the laser ranging reference value and the real-time vertical distance. The formula for calculating Δh is F = k1・Δh, where F is the contact force between the conveyor belt 11 and the conveyor belt support 13, and k1 is the stiffness coefficient of the elastic support component. The data fusion processing unit is used to correct the absorbance spectral value calculated from the spectral signal by using the instantaneous thickness of the object under test as the optical path length parameter based on the Beer-Lambert law, or to substitute the instantaneous thickness of the object under test into the internal quality quantitative prediction model to calculate the internal component content parameter of the object under test, so as to determine whether the object under test has defects.

[0109] In one embodiment, further reference is made to Figures 1 to 3 As shown, a maintenance cover 51 is provided on the spectral detection chamber 5. The maintenance cover 51 is installed on the spectral detection chamber 5 by an electrically controlled push rod 52, which controls the maintenance cover 51 to open automatically.

[0110] In one embodiment, further reference is made to Figures 1 to 3 As shown, an adjustable support leg 41 is also installed at the bottom of the frame 4. An adjusting screw 42 is installed on the adjustable support leg 41 to adjust the height of the adjustable support leg 41.

[0111] In one embodiment, the object being tested refers to a spherical or near-spherical agricultural product, but the present invention is not limited thereto.

[0112] The spectral detection system provided by this invention achieves efficient and stable transport and intelligent adjustment of the object under test through a V-shaped adaptive clamping and conveying device. This device employs a conveyor belt unit with a V-shaped spatial layout and a bottom-to-bottom transparent optical path penetration groove. Combined with an adjusting roller structure at the entrance, it automatically pushes the object under test to the center of the V-shaped angle upon entry, ensuring consistent detection position. Through the spatially coordinated layout of a dynamically adjustable transmission optical module and a non-contact laser ranging module, the light source component and detection component in the dynamically adjustable transmission optical module are located on the upper and lower sides of the optical path penetration groove, respectively, emitting and receiving spectral signals penetrating the object under test. The laser ranging unit, positioned adjacent to the optical path penetration groove, simultaneously performs non-contact ranging on the upper surface of the object under test. Combined with the lower surface reference displacement obtained by the pressure sensing reference mechanism, spatial synchronization of "spectral characteristic detection" and "diameter measurement" is achieved. This allows for the simultaneous acquisition of the object's intrinsic quality (spectral reflection) and external morphology (diameter) parameters in a single dynamic transport, overcoming the limitations of traditional single-dimensional detection. Furthermore, by coordinating the calibration of laser ranging and V-shaped transport reference, and by closely arranging the optical path penetration point, the protection of soft objects by non-contact measurement is preserved, while the high directionality and small spot characteristics of laser enable precise capture of the thickness of any point of the object under dynamic conditions, thus resolving the contradiction between the easy damage to objects by contact measurement and the insufficient accuracy of non-contact measurement.

[0113] Example 2

[0114] The following are method embodiments corresponding to the above-described spectral detection system embodiments, such as... Figure 6 As shown, Figure 6 A schematic diagram of the overall flow of the spectral detection method provided in Embodiment 2 of the present invention is shown. This method implementation can be implemented in conjunction with the above-described system implementation. The relevant technical details mentioned in the above system implementation remain valid in this method implementation, and will not be repeated here to avoid repetition.

[0115] A spectral detection method includes the following steps:

[0116] Step S1: Perform offline calibration. Obtain the vertical distance from the laser ranging unit's emitting end to the horizontal plane where the optical path penetration groove is located through the non-contact laser ranging diameter measurement module, and use it as the laser ranging reference value.

[0117] Before online detection, a one-time offline calibration and model training are required. When the adaptive clamping and conveying device is running unloaded, an obstacle is filled into the optical path penetration slot up to its upper surface. The vertical distance from the obstacle to the upper surface is detected by the laser ranging unit, and the vertical distance data is recorded. This vertical distance is the vertical distance from the laser ranging unit's transmitter to the horizontal plane where the optical path penetration slot is located, and this vertical distance is defined as the laser ranging reference value H0. Then, the obstacle is removed, and under the same unloaded conditions, the light source is turned on, and the corresponding spectral signal is collected through the dynamically adjustable transmission optical module as the reference light intensity I0(λ).

[0118] Step S2: Use an adaptive clamping and conveying device to transport the object under test so that the object under test passes through the detection area.

[0119] The object to be tested is placed at the entrance of the adaptive clamping and conveying device. Through two conveying units arranged in a V-shape and an adjustment structure, the object is pushed to the V-angle line between the two conveying units, automatically centering and straightening its posture so that its bottom adheres to the conveyor belt to form a stable reference. At this point, the control unit automatically adjusts the conveyor belt speed and the preload of the elastic element in the elastic support assembly based on the contact force detected by the pressure sensor. The preload is obtained by pressing a constant force spring through a lifting device, achieving adaptive switching between "light clamping of soft objects + high-speed conveying" and "stable clamping of hard objects + precise measurement."

[0120] Step S3: Acquire the spectral signal of the object under test through a dynamically adjustable transmission optical module.

[0121] Step S4: Collect the distance measurement data of the object being measured using a non-contact laser ranging diameter measurement module.

[0122] Step S5: Obtain the spectral signal and the ranging data through the data processing unit, calculate the thickness of the object under test based on the spectral signal and the ranging data, correct the spectral signal or predict the internal component content of the object under test based on the thickness, and generate a defect judgment result of the object under test.

[0123] In one embodiment, in step S4, the thickness is the instantaneous thickness D(t), which is calculated as follows: D(t) = H0 - H1(t) - Δh, where H0 is the laser ranging reference value, H1(t) is the real-time vertical distance collected by the non-contact laser ranging diameter measurement module, and Δh is the deformation of the elastic support component in the adaptive clamping and conveying device after being compressed. The ranging data includes the laser ranging reference value and the real-time vertical distance.

[0124] In one embodiment, step S5, correcting the spectral signal based on the thickness, includes:

[0125] Calculate the absorbance spectral value A(λ) = -log10(I(λ) / I0(λ)), where I(λ) is the real-time spectral signal transmitted through the object being tested, acquired by the detection component, and I0(λ) is the reference light intensity;

[0126] The instantaneous thickness of the object under test is used as the optical path length parameter to correct the absorbance spectral value. The correction formula is: A _corrected =k*A(λ) / D(t), where A _corrected The absorbance spectral value is the corrected value, and k is a proportionality constant or model coefficient.

[0127] In one embodiment, in step S5, predicting the internal component content of the object being tested includes:

[0128] Substitute the spectral signal and the thickness into the internal quality quantitative prediction model to output the internal component content of the measured object;

[0129] The quantitative prediction model for internal quality is trained through the following steps: preparing several test object samples covering different sizes and internal qualities; acquiring spectral signals and ranging data of each sample through the dynamically adjustable transmission optical module and the non-contact laser ranging diameter measurement module to obtain spectral-thickness data; determining the true value of the internal component content of each sample; using the spectral-thickness data as input variables and the true value of the internal component content as output variables, training the quantitative prediction model for internal quality through partial least squares regression, support vector machine algorithm, or deep learning algorithm.

[0130] In one embodiment, in step S5, generating the defect judgment result of the tested object includes:

[0131] Substitute the spectral signal and the thickness into the internal defect discrimination model to output the defect judgment result of the tested object;

[0132] The internal defect discrimination model is trained through the following steps:

[0133] Prepare several sample objects of varying sizes and internal qualities. Acquire spectral signals and ranging data for each sample using the dynamically adjustable transmission optics module and the non-contact laser ranging diameter measurement module to obtain spectral-thickness data. Manually determine the internal defect category of each sample, for example, 0 = no defect, 1 = defective, 1.1 = hollow, 1.2 = black core, 1.3 = ring corrosion, etc. Using the spectral-thickness data as input variables and the internal defect category as output variables, train an internal defect discrimination model using deep learning, logistic regression, or an SVM classifier.

[0134] In one embodiment, the internal defect discrimination model and the internal component content prediction model are deployed in the data fusion processing unit. The data fusion processing unit calculates the absorbance spectrum A(λ), takes the absorbance spectrum A(λ) and the corresponding maximum thickness Dmax as input, and simultaneously substitutes them into the deployed internal defect discrimination model and the internal component content prediction model to output the defect discrimination result of whether the tested object has defects and what component content the tested object has.

[0135] In one embodiment, the tested object refers to spherical or near-spherical agricultural products, but the present invention is not limited thereto. The internal component content prediction model mainly includes the determination of the content of components such as dry matter, starch, reducing sugar, protein, and vitamin C.

[0136] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims and not by the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A spectral detection system, characterized in that, include: An adaptive clamping and conveying device includes two conveying units arranged in a V-shaped angle along their conveying direction, and a light path penetration groove that is open at both the top and bottom is provided between the bottoms of the two conveying units. A dynamically adjustable transmission optical module includes a light source assembly and a detection assembly. The light source assembly and the detection assembly are respectively arranged on the upper and lower sides of the optical path penetration slot. The light source assembly is used to emit detection light rays that pass through the optical path penetration slot and the object under test. The detection assembly is used to receive the detection light rays and collect the spectral signal that penetrates the object under test. A non-contact laser ranging diameter measurement module includes a laser ranging unit, which is positioned facing the optical path penetration groove; The data processing unit is electrically connected to the dynamically adjustable transmission optical module and the non-contact laser ranging diameter measurement module, respectively. The control unit is connected to the adaptive clamping and conveying device, the dynamically adjustable transmission optical module, and the non-contact laser ranging and diameter measurement module, respectively.

2. The system as described in claim 1, characterized in that, The data processing unit is used to acquire the spectral signal and the ranging data collected by the non-contact laser ranging diameter measurement module, process the spectral signal and ranging data to calculate the thickness of the object under test, correct the spectral signal based on the thickness or predict the internal composition content of the object under test, and generate a defect judgment result of the object under test.

3. The system as described in claim 1, characterized in that, The data processing unit includes a synchronous data processing unit and a data fusion processing unit; The synchronous data processing unit is used to calculate the instantaneous thickness D(t) = H0 - H1(t) - Δh of the measured object based on the ranging data, where H0 is the vertical distance from the laser ranging unit transmitter to the horizontal plane where the optical path penetration slot is located, which is the laser ranging reference value; H1(t) is the real-time vertical distance from the laser ranging unit transmitter to the measured object; and Δh is the deformation of the elastic support component after being compressed. The ranging data includes the laser ranging reference value and the real-time vertical distance. The data fusion processing unit is used to correct the absorbance spectral value calculated from the spectral signal by using the instantaneous thickness of the object under test as an optical path length parameter based on the Beer-Lambert law, or to substitute the instantaneous thickness of the object under test into the internal quality quantitative prediction model to calculate the internal component content parameter of the object under test.

4. The system as described in claim 1, characterized in that, The adaptive clamping and conveying device also includes a conveyor belt support, the conveying unit is a conveyor belt, the conveyor belt is slidably arranged in a V-shape on the conveyor belt support, and the optical path penetration groove is opened at the bottom of the V-shape of the conveyor belt support.

5. The system as described in claim 4, characterized in that, The top of the conveyor belt support is rotatably equipped with an adjustment structure, which is located at the entrance of the adaptive clamping conveyor and is offset from the optical path penetration groove; and there is a gap between the bottom of the adjustment structure and the conveyor belt.

6. The system as described in claim 4, characterized in that, It also includes a pressure sensing reference mechanism, which includes an elastic support component and a pressure sensor. The two sides of the conveyor belt support away from the optical path penetration groove are mounted on the frame through the elastic support component, and the pressure sensor is disposed between the conveyor belt and the conveyor belt support.

7. The system as described in claim 6, characterized in that, The elastic support assembly includes a lifting device and a constant force spring. The constant force spring is disposed on the top of the lifting device, and the lifting device is controlled and connected to the control unit.

8. The system as described in claim 1, characterized in that, The adaptive clamping and conveying device is provided with a mounting beam above it, and a mounting frame is provided on the mounting beam. The laser ranging unit and the detection component are both mounted on the mounting frame. The mounting beam has a first adjustment groove extending in the left-right direction, and the mounting frame has a second adjustment groove extending in the up-down direction. The first adjustment groove and the second adjustment groove are connected by fastening bolts.

9. The system as described in claim 1, characterized in that, The detection component is provided with sample sensors at intervals on one side facing the feed inlet of the adaptive clamping conveyor, and the sample sensors are positioned facing the optical path penetration groove.

10. A spectral detection method, characterized in that, include: An adaptive clamping and conveying device is used to transport the object under test so that the object under test passes through the detection area; The spectral signal of the object under test is acquired by a dynamically adjustable transmission optical module. The distance measurement data of the object being measured is acquired by a non-contact laser ranging diameter measurement module; The data processing unit acquires the spectral signal and the ranging data, calculates the thickness of the object under test based on the spectral signal and the ranging data, corrects the spectral signal or predicts the internal component content of the object under test based on the thickness, and generates a defect judgment result for the object under test.

11. The method as described in claim 10, characterized in that, Before the step of using the adaptive clamping and conveying device to convey the object under test, the method further includes: Offline calibration is performed, and the vertical distance from the laser ranging unit's emitting end to the horizontal plane where the optical path penetration groove is located is obtained through the non-contact laser ranging diameter measurement module, which serves as the laser ranging reference value. The corresponding spectral signal is acquired by the dynamically adjustable transmission optical module as a reference light intensity.

12. The method as described in claim 11, characterized in that, The method of using an adaptive clamping and conveying device to transport the object under test includes: Place the object to be tested into the inlet of the adaptive clamping and conveying device. The object under test is pushed to the V-shaped angle line between the two transmission units and the adjustment structure through two transmission units arranged in a V-shape, thereby automatically centering and correcting the posture of the object under test.

13. The method as described in claim 11, characterized in that, The thickness is the instantaneous thickness D(t), which is calculated using the formula: D(t) = H0 - H1(t) - Δh, where H0 is the laser ranging reference value, H1(t) is the real-time vertical distance collected by the non-contact laser ranging diameter measurement module, and Δh is the deformation of the elastic support component in the adaptive clamping and conveying device after being compressed. The ranging data includes the laser ranging reference value and the real-time vertical distance.

14. The method as described in claim 13, characterized in that, Correcting the spectral signal based on the thickness includes: Calculate the absorbance spectral value A(λ) = -log10(I(λ) / I0(λ)), where I(λ) is the real-time spectral signal transmitted through the object being tested, acquired by the detection component, and I0(λ) is the reference light intensity; The instantaneous thickness of the object under test is used as the optical path length parameter to correct the absorbance spectral value. The correction formula is: A _corrected =k*A(λ) / D(t), where A _corrected The absorbance spectral value is the corrected value, and k is a proportionality constant or model coefficient.

15. The method as described in claim 12, characterized in that, The prediction of the internal component content of the object being measured includes: Substitute the spectral signal and the thickness into the internal quality quantitative prediction model to output the internal component content of the measured object; The internal quality quantitative prediction model is trained through the following steps: Prepare several test object samples covering different sizes and internal qualities; The spectral signals and ranging data of each sample are acquired by the dynamically adjustable transmission optical module and the non-contact laser ranging diameter measurement module to obtain spectral-thickness data. Determine the true values ​​of the internal component content of each sample; Using the spectral-thickness data as input variables and the true values ​​of the internal component content as output variables, the internal quality quantitative prediction model is trained using partial least squares regression, support vector machine algorithm, or deep learning algorithm.

16. The method as described in claim 12, characterized in that, The generation of the defect judgment result of the tested object includes: Substitute the spectral signal and the thickness into the internal defect discrimination model to output the defect judgment result of the tested object; The internal defect discrimination model is trained through the following steps: Prepare several test object samples covering different sizes and internal qualities; The spectral signals and ranging data of each sample are acquired by the dynamically adjustable transmission optical module and the non-contact laser ranging diameter measurement module to obtain spectral-thickness data. Manually determine the type of internal defect in each sample; Using the spectral-thickness data as input variables and the internal defect category as output variables, the internal defect discrimination model is trained using deep learning, logistic regression, or SVM classifiers.

Images