A Hyperspectral-Based Detection Device for the Cortex of Natural Rubber Trees and Its Detection Method

Abstract

The present invention relates to a detection device and a detection method for natural rubber trees based on hyperspectral technology. The device includes a walking chassis, a robotic arm, an imaging spectrometer, a vision camera, a portable support cloud platform, a laser scanning system, a data acquisition system, a data transmission module, a data processing module, and a power supply system. The imaging spectrometer is used to obtain the spectral reflection data of different layers of the rubber tree bark, combined with laser scanning to achieve efficient sampling, and the data is sent to the analysis terminal through wireless transmission. Through the built-in data processing module, spectral separation and feature extraction technologies are adopted, combined with machine learning models to achieve predictions of the bark health status, tree age, and pests and diseases. The device of the present invention has the characteristics of portability, non-destructiveness, high efficiency, and intelligence, and is suitable for tree health management in natural rubber plantations and precision agriculture monitoring.

Description

Technical Field

[0001] The present invention relates to the field of automated machinery, and particularly to a detection device and method for the cortex of natural rubber trees based on hyperspectral technology. Background Art

[0002] Natural rubber trees are important industrial plants and are widely used in rubber production. However, during the growth process of rubber trees, the health of the bark is affected by various factors such as pests and diseases, environmental factors, and growth stages. Traditional bark detection methods, such as manual inspection and chemical analysis, have significant limitations, and are cumbersome to operate and inefficient. Existing non-destructive detection methods are mostly based on visual images or near-infrared technology, but these methods often cannot efficiently and accurately analyze different tree cortexes of rubber trees. Especially in the case of complex bark layers and variable samples, it is difficult to ensure the accuracy and consistency of the detection results. Summary of the Invention

[0003] The main object of the present invention is to provide a detection device and method for the cortex of natural rubber trees based on hyperspectral technology, which uses hyperspectral imaging technology to accurately detect and analyze different layers of rubber tree bark, and solves the problems of difficult bark detection, low accuracy, and low efficiency in the prior art. Through the technical solution of the present invention, different tree cortexes of natural rubber trees can be detected efficiently and non-destructively, and relevant information such as health status, tree age, and tree species can be obtained, providing data support for agricultural management and rubber production.

[0004] To achieve the above object, the technical solution adopted by the present invention is as follows:

[0005] A detection device for the cortex of natural rubber trees based on hyperspectral technology includes a walking chassis, a robotic arm, and a hyperspectral detection platform. The robotic arm is fixedly installed on the walking chassis, and a hyperspectral detection platform is fixedly installed at the end of the robotic arm. The hyperspectral detection platform includes a laser scanning system and a portable support cloud platform. The laser scanning system includes an imaging spectrometer and a vision camera. The imaging spectrometer is fixedly installed on the hyperspectral detection platform, and the vision camera is fixedly installed on the robotic arm. The portable support cloud platform includes a platform lower bottom plate, a shock-absorbing rubber sleeve, a platform upper bottom plate, a longitudinal support arm, a horizontal support arm, a pitching motor, a transverse support arm, and a yaw motor. The platform lower bottom plate is fixedly installed at the end of the robotic arm, the shock-absorbing rubber sleeve is fixedly installed on the platform lower bottom plate, the platform upper bottom plate is fixedly installed above the shock-absorbing rubber sleeve, the longitudinal support arm is fixedly installed on the platform upper bottom plate, the pitching motor is fixedly installed on the longitudinal support arm, the output end of the pitching motor is fixedly connected to the transverse support arm, the yaw motor is fixedly installed on the transverse support arm, the output end of the yaw motor is fixedly connected to the horizontal support arm, and the imaging spectrometer is fixedly connected to the horizontal support arm.

[0006] Further, the laser scanning system further includes a light shield and a supplementary light source, and both the supplementary light source and the light shield are fixedly installed on the horizontal support arm.

[0007] Further, the robotic arm includes a first link body, a first link motor, a first link sector gear, a second link body, a second link motor, a second link sector gear, and a third link. The third link is fixedly installed on the walking chassis. The second link body is swingably connected to the third link. A second link sector gear is fixedly installed on the side end face of the connection end of the third link and the second link body. A second link motor is fixedly installed on the second link body. A second swing gear is fixedly connected to the output end of the second link motor. The second swing gear is meshed and connected with the second link sector gear. The other end of the second link body is swingably connected to the first link body. A first link sector gear is fixedly installed on the side end face of the connection of the second link body and the first link body. A first link motor is fixedly installed on the first link body. A first swing gear is fixedly connected to the output end of the first link motor. The first swing gear is meshed and connected with the first link sector gear.

[0008] Further, the walking chassis includes a front drive wheel system, a rear drive wheel system, a chassis main body, a crawler belt, and a power supply system. The front drive wheel system and the rear drive wheel system are respectively fixedly installed at the front end and the rear end of the chassis main body. A power supply system is also fixedly installed on the chassis main body. The front drive wheel system and the rear drive wheel system are connected by the crawler belt.

[0009] Further, the imaging spectrometer is used to obtain spectral reflection data of different layers of the bark of natural rubber trees, and the wavelength range covered by the imaging spectrometer is from 1000 nm to 2500 nm;

[0010] The main structure of the portable support cloud platform is made of lightweight and high-strength materials and is used to mount the imaging spectrometer to facilitate the stable operation of the imaging spectrometer when the device moves in the woods;

[0011] The laser scanning system is used to scan the surface of rubber trees and guide the imaging spectrometer to perform precise sampling on the target area.

[0012] Further, a data acquisition system, a data transmission module, and a data processing module are also fixedly installed on the walking chassis. The data acquisition system is electrically connected to the laser scanning system and the imaging spectrometer. The data acquisition system is electrically connected to the data transmission module. The data transmission module is electrically connected to the data processing module;

[0013] The data processing module performs noise reduction, spectral separation, and feature extraction on the collected spectral data to generate spectral features for analysis. The data acquisition system, the data transmission module, and the data processing module are all electrically connected to the power supply system.

[0014] Further, the laser scanning system can adjust the scanning range and resolution according to the texture characteristics of the rubber tree surface; both the data acquisition system and the data transmission module are equipped with Bluetooth modules and Wi-Fi modules for data transmission within short and long distances; the power supply system includes a solar-assisted power supply module for extending the battery life of the device during field operations; the data processing module uses the principal component analysis algorithm to reduce the dimensionality of the spectral data and combines the support vector machine model to classify and predict the health status, tree age, and diseases of the rubber tree bark; the imaging spectrometer works in cooperation with the supplementary light source to provide uniform illumination for the target bark area in case of insufficient light.

[0015] Further, a detection method for a portable hyperspectral natural rubber tree cortex detection device is characterized by comprising the following steps:

[0016] Step 1, the vision camera identifies the distance L1 between the detection device and the natural rubber tree, and the data processing module outputs the speeds ω1 of the front drive wheel system and ω2 of the rear drive wheel system of the control walking chassis according to the natural rubber forest environment;

[0017] Step 2, when the distance L1 identified and detected by the vision camera reaches the target distance range, the walking chassis stops moving, and the vision camera further identifies and detects the spatial positions x0, y0, and z0 of the bark to be detected of the target natural rubber tree, and feeds the information back to the data processing module through the data acquisition system and the data transmission module;

[0018] Step 3, the initial spatial positions of the end of the robotic arm are x1, y1, and z1, and the data processing module controls the end of the robotic arm to move to the target spatial positions x2, y2, and z2;

[0019] Step 4, the data processing module controls the laser scanning system of the hyperspectral detection platform to perform a surface scan on the rubber tree bark to identify the target area;

[0020] Step 5, use the imaging spectrometer to obtain the spectral reflection data of the rough cortex, sandy cortex, yellow cortex, and cyst cortex of the natural rubber tree bark, and transmit the spectral data to the data processing module through the data acquisition system and the data transmission module;

[0021] Step 6, in the data processing module, perform noise reduction, band selection, and feature extraction on the spectral data to establish a classification, detection, identification, and discrimination network model for different cortices of natural rubber trees;

[0022] Step 7, use a machine learning-based model to analyze the spectral data and generate a bark health report, including tree age estimation, health status diagnosis, and pest and disease warning information.

[0023] Further, in step four, wavelet transform technology is used to denoise the spectral data to improve the accuracy of data analysis; in step five, the machine learning model combines deep learning algorithms, and through multi-layer convolution of spectral data of different tree barks, healthy features are extracted and the classification results are optimized.

[0024] Further, the bark health report is displayed through a graphical interface on a mobile terminal, and supports users to view the detection data and analysis results in real time.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0026] The present invention uses hyperspectral imaging technology to accurately detect and analyze different layers of rubber tree bark, and solves the problems of difficult bark detection, low accuracy, and low efficiency in the prior art. Through the technical solution of the present invention, different tree bark layers of natural rubber trees can be detected efficiently and non-destructively, and relevant information such as health status, tree age, and tree species can be obtained, providing data support for agricultural management and rubber production. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a three-dimensional structure schematic diagram of the present invention;

[0028] Figure 2 is a partial three-dimensional structure schematic diagram of the portable support cloud platform of the present invention;

[0029] Figure 3 is a schematic diagram of the control flow of the present invention;

[0030] Figure 4 is a schematic diagram of the working flow of the present invention.

[0031] Reference Signs:

[0032] 100. Hyperspectral detection platform; 200. Robot arm; 300. Walking chassis; 101. Lower platform base plate; 102. Shock-absorbing rubber sleeve; 103. Upper platform base plate; 104. Longitudinal support arm; 105. Light-shielding cover; 106. Supplementary light source; 107. Transverse support arm; 108. Pitch motor; 109. Horizontal support arm; 110. Yaw motor; 111. Imaging spectrometer; 112. Vision camera; 201. First link body; 202. First link motor; 203. First link sector gear; 204. Second link body; 205. Second link motor; 206. Second link sector gear; 207. Third link; 301. Front drive wheel system; 302. Rear drive wheel system; 303. Chassis main body; 304. Track; 600. Power supply system; 400. Data transmission module; 500. Data processing module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] The following elaborates on the preferred embodiments of the present invention in conjunction with the accompanying drawings, so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making the protection scope of the present invention more clearly defined.

[0034] Refer to Figures 1-4 As shown, a hyperspectral-based natural rubber tree cortex detection device includes a walking chassis 300, a robotic arm 200, and a hyperspectral detection platform 100. The robotic arm 200 is fixedly installed on the walking chassis 300, and the hyperspectral detection platform 100 is fixedly installed at the end of the robotic arm 200. The hyperspectral detection platform 100 includes a laser scanning system and a portable support cloud platform. The laser scanning system includes an imaging spectrometer 111 and a vision camera 112. The imaging spectrometer 111 is fixedly installed on the hyperspectral detection platform 100, and the vision camera 112 is fixedly installed on the robotic arm 200. The portable support cloud platform includes a platform lower base plate 101, a shock-absorbing rubber sleeve 102, a platform upper base plate 103, a longitudinal support arm 104, a horizontal support arm 109, a pitching motor 108, a transverse support arm 107, and a yaw motor 110. The platform lower base plate 101 is fixedly installed at the end of the robotic arm 200. A shock-absorbing rubber sleeve is fixedly installed on the platform lower base plate 101. The platform upper base plate 103 is fixedly installed above the shock-absorbing rubber sleeve 102. The shock-absorbing rubber sleeve 102 is used for the overall earthquake protection of the platform. The longitudinal support arm 104 is fixedly installed on the platform upper base plate 103. The pitching motor 108 is fixedly installed on the longitudinal support arm 104. The output end of the pitching motor 108 is fixedly connected to the transverse support arm 107. The yaw motor 110 is fixedly installed on the transverse support arm 107. The output end of the yaw motor 110 is fixedly connected to the horizontal support arm 109. The imaging spectrometer 111 is fixedly connected to the horizontal support arm 109. The imaging spectrometer 111 is fixed by the horizontal support arm 109. The yaw motor 110 is used to adjust the clockwise / counterclockwise rotation angle of the imaging spectrometer 111. The transverse support arm 107 is used to support the yaw motor 110. The pitching motor 108 is used to adjust the pitching angle of the imaging spectrometer.

[0035] The laser scanning system further includes a light shield 105 and a supplementary light source 106. Both the supplementary light source 106 and the light shield 105 are fixedly installed on the horizontal support arm 109.

[0036] The robotic arm 200 includes a first link body 201, a first link motor 202, a first link sector gear 203, a second link body 204, a second link motor 205, a second link sector gear 206, and a third link 207. The third link 207 is fixedly installed on the mobile chassis 300. The second link body 204 is swingably connected to the third link 207. A second link sector gear 206 is fixedly installed on the side end face of the connection end of the third link 207 and the second link body 204. A second link motor 205 is fixedly installed on the second link body 204. A second swing gear is fixedly connected to the output end of the second link motor 205. The second swing gear is meshed and connected with the second link sector gear 206. The other end of the second link body 204 is swingably connected to the first link body 201. A first link sector gear 203 is fixedly installed on the side end face of the connection between the second link body 204 and the first link body 201. A first link motor 202 is fixedly installed on the first link body 201. A first swing gear is fixedly connected to the output end of the first link motor 202. The first swing gear is meshed with the first link sector gear.

[0037] The mobile chassis 300 includes a front drive wheel system 301, a rear drive wheel system 302, a chassis main body 303, a crawler 304, and a power supply system 600. The front drive wheel system 301 and the rear drive wheel system 302 are respectively fixedly installed at the front end and the rear end of the chassis main body 303. A power supply system 600 is also fixedly installed on the chassis main body 303. The front drive wheel system 301 and the rear drive wheel system 302 are connected by a crawler 304.

[0038] The imaging spectrometer 111 is used to obtain spectral reflection data of different layers of the bark of natural rubber trees. The wavelength range covered by the imaging spectrometer 111 is from 1000 nm to 2500 nm;

[0039] The main structure of the portable support pan-tilt is made of lightweight and high-strength materials and is used to mount the imaging spectrometer 111 to facilitate the stable operation of the imaging spectrometer 111 when the device moves in the forest;

[0040] The laser scanning system is used to scan the surface of the rubber tree and guide the imaging spectrometer 111 to perform precise sampling on the target area.

[0041] A data acquisition system, a data transmission module 400, and a data processing module 500 are also fixedly installed on the mobile chassis 300. The data acquisition system is electrically connected to the laser scanning system and the imaging spectrometer 111. The data acquisition system is electrically connected to the data transmission module 400. The data transmission module 400 is electrically connected to the data processing module 500;

[0042] The data processing module 500 performs noise reduction, spectral separation, and feature extraction on the collected spectral data to generate spectral features for analysis. In this data acquisition system, the data transmission module 400 and the data processing module 500 are both electrically connected to the power supply system 600.

[0043] The laser scanning system can adjust the scanning range and resolution according to the texture characteristics of the rubber tree surface; both the data acquisition system and the data transmission module 400 are equipped with Bluetooth modules and Wi-Fi modules for data transmission within short and long distances; the power supply system 600 includes a solar-assisted power supply module to extend the endurance time of the device during field operations; the data processing module 500 uses the principal component analysis (PCA) algorithm to reduce the dimensionality of the spectral data and combines it with the support vector machine (SVM) model to classify and predict the health status, tree age, and diseases of the rubber tree bark; the imaging spectrometer 111 and the supplementary light source 106 work together to provide uniform illumination to the target bark area in case of insufficient light.

[0044] The detection method includes the following steps:

[0045] Step 1: The vision camera 112 identifies the distance L1 between the detection device and the natural rubber tree, and the data processing module 500 outputs the speeds ω1 of the front drive wheel system 301 and ω2 of the rear drive wheel system 302 of the control walking chassis 300 according to the natural rubber forest environment.

[0046] Step 2: When the distance L1 identified and detected by the vision camera 112 reaches the target distance range, the walking chassis 300 stops moving. The vision camera further identifies and detects the spatial positions x0, y0, and z0 of the bark to be detected of the target natural rubber tree, and feeds the information back to the data processing module 500 through the data acquisition system and the data transmission module 400.

[0047] Step 3: The initial spatial positions of the end of the robotic arm 200 are x1, y1, and z1, and the data processing module 500 controls the end of the robotic arm 200 to move to the target spatial positions x2, y2, and z2.

[0048] Step 4: The data processing module 500 controls the laser scanning system of the hyperspectral detection platform 100 to perform a surface scan on the rubber tree bark to identify the target area.

[0049] Step 5: Use the imaging spectrometer 111 to obtain the spectral reflection data of the rough cortex, sandy cortex, yellow cortex, and cyst cortex of the natural rubber tree bark, and transmit the spectral data to the data processing module 500 through the data acquisition system and the data transmission module 400.

[0050] Step Six: In the data processing module 500, denoise the spectral data, select bands, and extract features, and establish a classification, detection, recognition, and discrimination network model for different cortexes of natural rubber trees;

[0051] Step Seven: Use a machine learning-based model to analyze the spectral data and generate a bark health report, including tree age estimation, health condition diagnosis, and pest and disease warning information.

[0052] In Step Four, wavelet transform technology is used to denoise the spectral data to improve the accuracy of data analysis; in Step Five, the machine learning model combines deep learning algorithms, and through multi-layer convolution of the spectral data of different tree cortexes, health features are extracted and the classification results are optimized.

[0053] The bark health report is displayed through a graphical interface on a mobile terminal, and supports users to view the detection data and analysis results in real time.

[0054] The above are only the preferred embodiments of the present invention, and the protection scope of the present invention is not limited to the above embodiments. Any equivalent modifications or changes made by those of ordinary skill in the art according to the disclosed content of the present invention shall be included in the protection scope recorded in the claims.

Claims

1. A natural rubber tree cortex layer detection device based on hyperspectral, comprising a walking chassis (300), a mechanical arm (200) and a hyperspectral detection platform (100), wherein the walking chassis (300) is fixedly mounted with the mechanical arm (200), and the end of the mechanical arm (200) is fixedly mounted with the hyperspectral detection platform (100), characterized in that: The hyperspectral detection platform (100) comprises a laser scanning system and a portable support platform. The laser scanning system comprises an imaging spectrometer (111) and a visual camera (112). The imaging spectrometer (111) is fixedly mounted on the hyperspectral detection platform (100). The visual camera (112) is fixedly mounted on a mechanical arm (200). The portable support platform comprises a platform bottom plate (101), a shock-absorbing rubber sleeve (102), a platform bottom plate (103), a longitudinal support arm (104), a horizontal support arm (109), a pitch motor (108), a lateral support arm (107) and a yaw motor (110). The platform bottom plate (101) is fixedly mounted on the mechanical arm (200). 00), a shock-absorbing rubber sleeve (102) is fixedly installed on the lower bottom plate (101) of the platform, a platform upper bottom plate (103) is fixedly installed above the shock-absorbing rubber sleeve (102), a longitudinal support arm (104) is fixedly installed on the platform bottom plate (103), a pitch motor (108) is fixedly installed on the longitudinal support arm (104), a transverse support arm (107) is fixedly connected to the output end of the pitch motor (108), a yaw motor (110) is fixedly installed on the transverse support arm (107), the output end of the yaw motor (110) is fixedly connected to the horizontal support arm (109), and the horizontal support arm (109) is fixedly connected to the imaging spectrometer (111); The laser scanning system further comprises a light shield (105) and a supplementary light source (106), wherein the supplementary light source (106) and the light shield (105) are both fixedly mounted on the horizontal support arm (109); The mechanical arm (200) comprises a first connecting rod body (201), a first connecting rod motor (202), a first connecting rod sector gear (203), a second connecting rod body (204), a second connecting rod motor (205), a second connecting rod sector gear (206) and a third connecting rod (207). The third connecting rod (207) is fixedly mounted on a walking chassis (300). The third connecting rod (207) is swingably connected to the second connecting rod body (204). The second connecting rod sector gear (206) is fixedly mounted on the side end surface of the connecting end of the third connecting rod (207) and the second connecting rod body (204). The second connecting rod is fixedly mounted on the second connecting rod body (204). The motor (205) includes a second swing gear fixedly connected to the output end of the second connecting rod motor (205), the second swing gear is meshingly connected to the second connecting rod sector gear (206), the other end of the second connecting rod body (204) is swingably connected to the first connecting rod body (201), the side end surface where the second connecting rod body (204) is connected to the first connecting rod body (201) is fixedly mounted with the first connecting rod sector gear (203), the first connecting rod motor (202) is fixedly mounted on the first connecting rod body (201), the output end of the first connecting rod motor (202) is fixedly connected with the first swing gear, and the first swing gear is meshingly connected to the first connecting rod sector gear.

2. A natural rubber tree cortex layer detection device based on hyperspectral according to claim 1, characterized in that: The walking chassis (300) comprises a front driving wheel system (301), a rear driving wheel system (302), a chassis body (303), a crawler track (304) and a power supply system (600). The front driving wheel system (301) and the rear driving wheel system (302) are fixedly mounted at the front end and the rear end of the chassis body (303), respectively. The power supply system (600) is also fixedly mounted on the chassis body (303). The front driving wheel system (301) and the rear driving wheel system (302) are connected via the crawler track (304).

3. A natural rubber tree cortex layer detection device based on hyperspectral according to claim 1, characterized in that: The imaging spectrometer (111) is used to obtain spectral reflectance data of different layers of natural rubber tree bark, and the imaging spectrometer (111) covers a wavelength range of 1000nm to 2500nm; The main structure of the portable supporting platform is made of lightweight high-strength material and is used to mount the imaging spectrometer (111), so as to facilitate the stable operation of the imaging spectrometer (111) when the device is moved among the trees. The laser scanning system is used to scan the surface of the rubber tree and guide the imaging spectrometer (111) to accurately sample the target area.

4. The natural rubber tree cortex layer detection device based on hyperspectral according to claim 1, characterized in that: A data acquisition system, a data transmission module (400) and a data processing module (500) are also fixedly mounted on the walking chassis (300); the data acquisition system is electrically connected to the laser scanning system and the imaging spectrometer (111); the data acquisition system is electrically connected to the data transmission module (400); and the data transmission module (400) is electrically connected to the data processing module (500); The data processing module (500) performs noise reduction, spectrum separation and feature extraction on the collected spectral data to generate spectral features for analysis. The data collection system, the data transmission module (400) and the data processing module (500) are all electrically connected to the power supply system (600).

5. The natural rubber tree cortex layer detection device based on hyperspectral according to claim 4, characterized in that: The laser scanning system can adjust the scanning range and resolution according to the texture characteristics of the rubber tree surface; the data acquisition system and the data transmission module (400) are both provided with a Bluetooth module and a Wi-Fi module for realizing data transmission within a short distance and a long distance; the power supply system (600) includes a solar auxiliary power supply module for extending the endurance of the device in field operations; the data processing module (500) uses a principal component analysis (PCA) algorithm to reduce the dimension of the spectral data, and combines a support vector machine (SVM) model to classify and predict the health status, tree age and disease of the rubber tree bark; the imaging spectrometer (111) cooperates with the supplementary light source (106) to provide uniform illumination to the target bark area when the illumination is insufficient.

6. A detection method applied to the device according to any one of claims 1 to 5, characterized in that: The steps include: Step 1: The visual camera (112) identifies the distance L1 between the detection device and the natural rubber tree, and the data processing module (500) controls the speed ω1 of the front driving wheel system (301) and the speed ω2 of the rear driving wheel system (302) of the walking chassis (300) according to the natural rubber forest environment output; Step 2: When the distance L1 detected by the visual camera (112) reaches the target distance range, the walking chassis (300) stops moving, and the visual camera further identifies the spatial position of the bark of the natural rubber tree to be detected as x0, y0, z0, and feeds back the information to the data processing module (500) through the data acquisition system and the data transmission module (400); Step 3: The initial spatial position of the end of the robotic arm (200) is x1, y1, z1, and the data processing module (500) controls the end of the robotic arm (200) to move to the target spatial position x2, y2, z2; Step 4: The data processing module (500) controls the laser scanning system of the hyperspectral detection platform (100) to perform surface scanning on the rubber tree bark to identify the target area; Step 5: using an imaging spectrometer (111) to obtain spectral reflectance data of the coarse cortex, sand cortex, yellow cortex and cystic cortex of the natural rubber tree bark, and transmitting the spectral data to the data processing module (500) through a data acquisition system and a data transmission module (400); Step 6: In the data processing module (500), the spectral data is subjected to noise reduction, band selection and feature extraction, and a network model for classifying, detecting, identifying and distinguishing different cortical layers of natural rubber trees is established; Step 7: Use a machine learning-based model to analyze the spectral data and generate a bark health report, including tree age estimation, health status diagnosis, and pest and disease warning information.

7. The detection method of the device according to claim 6, characterized in that: In the step 4, the wavelet transform technology is used to reduce the noise of the spectral data to improve the accuracy of data analysis; in the step 5, the machine learning model is combined with the deep learning algorithm to extract health features and optimize the classification results through multi-layer convolution of the spectral data of different bark layers.

8. The detection method of the device according to claim 6, characterized in that: The bark health report is displayed in a graphical interface through a mobile terminal, and supports users to view test data and analysis results in real time.

Images