Movable crop stress in-vivo identification device and stress type identification method

Abstract

The invention provides a movable crop stress in-vivo recognition device and a stress type recognition method, and the device comprises a fixed support which is provided with a supporting rod extending in the transverse direction; the transverse displacement rod extends in the longitudinal direction, is arranged on the supporting rod in a sliding mode and is used for moving on the supporting rod in the transverse direction; the longitudinal displacement platform is arranged on the transverse displacement rod in a sliding mode and used for moving on the transverse displacement rod in the longitudinal direction; and the Raman spectrum acquisition assembly is arranged on the longitudinal displacement platform and is used for acquiring a Raman spectrum on the target plant. According to the movable crop stress in-vivo recognition device and the stress type recognition method provided by the invention, the advantages of no need of sample pretreatment and high specificity of a Raman spectrum technology are utilized, stress response marker molecules in plant tissues are detected, and related information of a plant physiological state is obtained; and in-vivo, lossless and rapid identification of crop stress is realized by matching with structures such as a fixed bracket, a transverse displacement rod and a longitudinal displacement platform.

Description

Technical Field

[0001] The present invention relates to the technical field of crop stress identification, and in particular to a movable crop stress on-body identification device and a stress type identification method. Background Art

[0002] When facing biotic or abiotic stress, crops will adapt to adverse environments through a series of physiological and morphological adjustments to maintain survival. However, these adaptive changes will cause plant growth to slow down, which will not only reduce crop yields, but may also affect product quality. Non-destructive and rapid identification of crop stress is crucial for taking timely stress resistance measures and ensuring agricultural production.

[0003] Traditional crop stress identification mainly relies on manual observation, which is inefficient, affected by human factors and has poor reliability. Laboratory detection methods based on physical and chemical analysis are complicated and time-consuming, and require sample collection, which causes damage to the plants and cannot achieve real-time monitoring. The common limitation of detection methods based on phenotypic imaging technology is that these methods focus on changes in plants at the morphological and physiological levels, and early stress identification has low sensitivity and poor accuracy.

[0004] Raman spectroscopy is often referred to as a "molecular fingerprint". It can detect changes in the content of stress marker molecules in plant tissues. These changes at the molecular level are the fastest and most accurate stress responses. Currently, Raman spectroscopy has been used to identify plant stresses. However, most applications use large laboratory Raman equipment, which makes it difficult to achieve in vivo detection. Detection based on handheld Raman equipment relies on manual and professional operations, which is inefficient and unstable for field applications. Summary of the invention

[0005] The invention provides a movable crop stress on-body identification device and a stress type identification method, which are used to solve the problems of low efficiency and poor stability of the existing detection method.

[0006] The present invention provides a movable crop stress on-body recognition device, comprising: A fixed bracket formed with a support rod extending in a transverse direction; A transverse displacement rod extending in the longitudinal direction and slidably disposed on the support rod for transverse movement on the support rod; A longitudinal displacement platform, slidably disposed on the transverse displacement rod, and configured to move longitudinally on the transverse displacement rod; The Raman spectrum acquisition component is arranged on the longitudinal displacement platform and is used to collect Raman spectra on the target plant, screen the spectral characteristic bands containing stress information, establish an abiotic stress classification model based on the spectral data of the band where the characteristic peak of the stress marker is located, and identify the stress type.

[0007] According to a movable crop stress on-body identification device provided by the present invention, the Raman spectrum acquisition component comprises: a leaf bin, a biaxial displacement mechanism and a Raman spectrometer; The leaf bin is arranged on the longitudinal displacement platform, and is formed with a leaf cavity, and the leaf cavity is used to accommodate leaves of the target plant; The detection end of the Raman spectrometer faces the blade cavity and is connected to the biaxial displacement mechanism, so that the position of the detection end of the Raman spectrometer facing the blade can be adjusted through the biaxial displacement mechanism.

[0008] According to a movable crop stress on-body recognition device provided by the present invention, the dual-axis displacement mechanism comprises: a movable platform, an X-axis motion component and a Y-axis motion component; The detection end of the Raman spectrometer is arranged on the movable platform, the movable platform is arranged on the X-axis moving component, and the X-axis moving component is used to drive the movable platform in a first lateral direction; the X-axis moving component is arranged on the Y-axis moving component, and the Y-axis moving component is used to drive the movable platform and the X-axis moving component in a second lateral direction.

[0009] According to a movable crop stress on-body identification device provided by the present invention, the leaf bin comprises: a bin door and a bin body, The bin door is rotatably arranged on the bin body, and the blade cavity with an opening on one side is formed between the bin door and the bin body.

[0010] According to the movable crop stress on-body identification device provided by the present invention, the blade bin is made of black material.

[0011] According to a movable crop stress on-body identification device provided by the present invention, the leaf bin further comprises: A blade clamp is connected to the compartment door and is used for clamping the blades arranged in the blade cavity.

[0012] According to a movable crop stress on-body identification device provided by the present invention, a plurality of blade clamps are provided, and the plurality of blade clamps are connected to the bin door at intervals.

[0013] According to a movable crop stress on-body identification device provided by the present invention, two support rods extending in the lateral direction are formed on the fixed bracket; One end of the lateral displacement rod can be slidably mounted on one of the support rods, and the other end of the lateral displacement rod can be slidably mounted on the other support rod.

[0014] According to the movable crop stress on-body identification device provided by the present invention, the longitudinal displacement platform can be slidably mounted on the transverse displacement rod.

[0015] The present invention also provides a method for identifying the type of stress using a movable crop stress on-body identification device, comprising: According to the spatial position of the target plant, the lateral displacement rod and the longitudinal displacement platform are moved so that the Raman spectrum collection component is located at the leaf position of the target plant; The leaves of the target plant are fixed to the leaf chamber, and the Raman spectrometer is moved by a dual-axis displacement mechanism to collect Raman spectra at multiple sampling points on the leaves; Preprocess the collected Raman spectral data and select the spectral characteristic bands with molecular information of stress markers in leaves; An abiotic stress classification model was established based on the spectral characteristic bands containing molecular information of stress markers in leaves to identify the types of stress.

[0016] The movable in-vivo crop stress identification device and stress type identification method provided by the present invention utilize the advantages of Raman spectroscopy technology that does not require sample pretreatment and has high specificity to detect stress response marker molecules in plant tissues and obtain information related to the physiological state of the plant. In combination with structures such as a fixed bracket, a lateral displacement rod and a longitudinal displacement platform, in-vivo, non-destructive and rapid identification of crop stress is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 It is a structural schematic diagram of the movable crop stress on-body identification device provided by the present invention.

[0019] Figure 2 It is a schematic diagram of the structure of the Raman spectrum collection component provided by the present invention.

[0020] Figure 3 It is a structural schematic diagram of the blade bin provided by the present invention.

[0021] Figure 4 It is a flow chart of a method for identifying the type of stress using a movable crop stress on-body identification device provided by the present invention.

[0022] Figure 5 Raman spectrum and characteristic peaks of corn leaves after data preprocessing.

[0023] Figure 6 Model results for classification of healthy, cold-stressed, and drought-stressed maize.

[0024] Reference numerals: 10. Fixed bracket; 100. Support rod; 20. Lateral displacement rod; 30. Longitudinal displacement platform; 40. Raman spectrum acquisition component; 410. Blade bin; 4110. Bin door; 4120. Bin body; 4130. Blade clamp; 420. Raman spectrometer; 430. Movable platform; 440. X-axis moving part; 450. Y-axis moving part; 50. Target plant; 60. Blade. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0026] Combine the following Figure 1-Figure 6 The present invention describes a movable crop stress on-body identification device and a stress type identification method.

[0027] The present application embodiment provides a movable crop stress on-body recognition device, such as Figures 1 to 3 As shown, the movable crop stress on-body identification device comprises: a fixed support 10, a transverse displacement rod 20, a longitudinal displacement platform 30 and a Raman spectrum collection component 40. The fixed support 10 is formed with a support rod 100 extending in the transverse direction; the transverse displacement rod 20 extends in the longitudinal direction and is slidably arranged on the support rod 100 for moving in the transverse direction on the support rod 100; the longitudinal displacement platform 30 is slidably arranged on the transverse displacement rod 20 for moving in the longitudinal direction on the transverse displacement rod 20; the Raman spectrum collection component 40 is arranged on the longitudinal displacement platform 30, and is used to collect Raman spectra on the target plant 50, screen the spectral characteristic bands containing stress information, so as to establish an abiotic stress classification model according to the spectral data of the band where the characteristic peak of the stress marker is located, and identify the type of stress.

[0028] In this embodiment, the support rod 100 on the fixed bracket 10 provides a sliding track for the lateral displacement rod 20. The lateral displacement rod 20 extends in the longitudinal direction (i.e., a direction perpendicular to the support rod 100) and is slidably disposed on the support rod 100. By adjusting the position of the lateral displacement rod 20, lateral movement on the support rod 100 can be achieved, thereby adjusting the lateral position of the Raman spectrum acquisition assembly 40.

[0029] The longitudinal displacement platform 30 is slidably disposed on the transverse displacement rod 20, and is used to move along the longitudinal direction (i.e., the direction parallel to the transverse displacement rod 20) on the transverse displacement rod 20. By adjusting the position of the longitudinal displacement platform 30, the longitudinal position of the Raman spectrum collection component 40 can be further fine-tuned to ensure that it can accurately reach the designated position of the target plant 50.

[0030] The Raman spectrum acquisition component 40 is arranged on the longitudinal displacement platform 30. It is responsible for collecting Raman spectra on the target plant 50, and by screening the spectral characteristic bands containing stress information, an abiotic stress classification model is established according to the spectral data of the band where the characteristic peak of the stress marker is located. This model can realize the accurate identification of the stress type.

[0031] When using the mobile crop stress on-body identification device of the embodiment of the present application, it is first necessary to accurately locate the target plant 50 according to its spatial position. This includes adjusting the positions of the lateral displacement rod 20 and the longitudinal displacement platform 30 so that the Raman spectrum acquisition component 40 can accurately reach the position of the leaf 60 of the target plant 50. The specific position of the leaf 60 of the target plant 50 is determined by naked eye observation or using other auxiliary equipment (such as a camera).

[0032] According to the lateral position of the target plant 50, the position of the lateral displacement rod 20 is slid on the support rod 100 to ensure that the lateral position of the Raman spectrum collection component 40 matches the target leaf 60. After the lateral displacement rod 20 is determined, the longitudinal position of the Raman spectrum collection component 40 is further fine-tuned by sliding the longitudinal displacement platform 30 on the lateral displacement rod 20 so that it can accurately reach the top of the target leaf 60.

[0033] Before the Raman spectrum is collected, the target leaf 60 needs to be fixed to prevent the accuracy of the spectrum data from being affected by the shaking of the leaf 60 during the collection process. After the leaf 60 is fixed, the Raman spectrum collection component 40 is used to collect the spectrum at multiple sampling points on the leaf 60. These sampling points should be evenly distributed on the leaf 60 to ensure that the collected spectrum data can fully reflect the stress state of the leaf 60.

[0034] The collected Raman spectral data are preprocessed, including noise removal, baseline correction, smoothing, etc., to improve the signal-to-noise ratio and accuracy of the spectral data. From the preprocessed spectral data, a spectral characteristic band having molecular information of stress markers in the leaf 60 is selected.

[0035] According to the selected spectral feature bands, an abiotic stress classification model is established using machine learning or deep learning algorithms. This model can automatically identify and classify the types of crop stress based on the input spectral feature band data. The preprocessed spectral feature band data is input into the established classification model, and the model automatically outputs the identification results of the stress type based on the input data. This result can help agricultural producers understand and respond to the stress status of crops in a timely manner, so as to take corresponding management measures to improve the yield and quality of crops.

[0036] The movable in vivo crop stress identification device provided in the embodiment of the present application utilizes the advantages of Raman spectroscopy technology, such as no need for sample pretreatment and high specificity, to detect stress response marker molecules in plant tissues and obtain information related to the physiological state of the plant. In conjunction with structures such as a fixed bracket 10, a lateral displacement rod 20, and a longitudinal displacement platform 30, in vivo, non-destructive, and rapid identification of crop stress is achieved.

[0037] In some embodiments, Figures 1 to 3 As shown, the Raman spectrum acquisition assembly 40 includes: a leaf bin 410, a biaxial displacement mechanism and a Raman spectrometer 420. The leaf bin 410 is arranged on the longitudinal displacement platform 30, and is formed with a leaf cavity, which is used to accommodate the leaf 60 of the target plant 50; the detection end of the Raman spectrometer 420 faces the leaf cavity, and is connected to the biaxial displacement mechanism, so as to adjust the position of the detection end of the Raman spectrometer 420 toward the leaf 60 through the biaxial displacement mechanism.

[0038] In this embodiment, the leaf chamber 410 is disposed on the longitudinal displacement platform 30 and forms a leaf chamber for accommodating the leaf 60 of the target plant 50. The design of the leaf chamber allows the leaf 60 to remain stable therein, preventing the leaf 60 from shaking during the spectrum collection process and affecting the accuracy of the data. In addition, the material and shape of the leaf chamber 410 are also carefully selected and designed to ensure that it has minimal interference with the Raman spectrum signal, thereby improving the quality of the spectrum data.

[0039] The dual-axis displacement mechanism allows the detection end of the Raman spectrometer 420 to be fine-tuned in both horizontal and vertical directions to ensure that the detection end can accurately align with the target sampling point on the leaf 60. This fine-tuning capability is critical to ensuring the accuracy and consistency of spectral data, especially when dealing with plants with complex leaf 60 structures.

[0040] The Raman spectrometer 420 is responsible for collecting Raman spectrum data of the leaf 60. The detection end of the Raman spectrometer 420 is connected to the leaf cavity through a dual-axis displacement mechanism, and can accurately locate the target sampling point on the leaf 60. When collecting spectral data, the Raman spectrometer 420 will emit a laser beam to the leaf 60 and collect the scattered light reflected by the leaf 60. By analyzing the wavelength and intensity of these scattered lights, the information of the stress marker molecules in the leaf 60 can be obtained, thereby realizing the identification of the crop stress state.

[0041] In practical applications, when it is necessary to collect the Raman spectrum of the leaf 60 of the target plant 50, the positions of the lateral displacement rod 20 and the longitudinal displacement platform 30 are first adjusted so that the leaf cavity of the leaf bin 410 can accommodate the target leaf 60. Then, the detection end position of the Raman spectrometer 420 is fine-tuned by the biaxial displacement mechanism to ensure that it can accurately align with the target sampling point on the leaf 60. Finally, the Raman spectrometer 420 is started to collect the spectrum, and the collected data is preprocessed and analyzed to obtain the information of the stress marker molecules in the leaf 60.

[0042] In some embodiments, Figures 1 to 3 As shown, the dual-axis displacement mechanism includes: a movable platform 430, an X-axis moving component 440 and a Y-axis moving component 450; the detection end of the Raman spectrometer 420 is arranged on the movable platform 430, the movable platform 430 is arranged on the X-axis moving component 440, and the X-axis moving component 440 is used to drive the movable platform 430 in a first lateral direction; the X-axis moving component 440 is arranged on the Y-axis moving component 450, and the Y-axis moving component 450 is used to drive the movable platform 430 and the X-axis moving component 440 in a second lateral direction.

[0043] In this embodiment, the movable platform 430 carries the detection end of the Raman spectrometer 420. The X-axis motion component 440 generally includes components such as a motor, a transmission mechanism, and a guide rail. The motor provides power, and transmits the power to the guide rail through the transmission mechanism, thereby driving the movable platform 430 to move along the guide rail in the X-axis direction. The design of the X-axis motion component 440 needs to ensure the smoothness and accuracy of the movement to meet the demand for precise control of the detection end of the Raman spectrometer 420. The Y-axis motion component 450 is similar to the X-axis motion component 440, and is used to drive the movable platform 430 (and the X-axis motion component 440) to move in the second lateral direction. The Y-axis motion component 450 also includes components such as a motor, a transmission mechanism, and a guide rail. Driven by the Y-axis motion component 450, the movable platform 430 can move along the Y-axis direction, thereby achieving fine adjustment of the detection end of the Raman spectrometer 420 in the vertical direction.

[0044] In practical applications, the working principle of the dual-axis displacement mechanism is as follows: when the position of the detection end of the Raman spectrometer 420 needs to be adjusted, the movable platform 430 is first driven to move in the X-axis direction by the X-axis motion component 440, so that the detection end is close to the sampling point of the target blade 60. Then, the movable platform 430 (and the X-axis motion component 440) are driven to move in the Y-axis direction by the Y-axis motion component 450, and the detection end is fine-tuned to ensure that it can accurately align with the target sampling point. The design of this dual-axis displacement mechanism enables the detection end of the Raman spectrometer 420 to achieve precise position adjustment in two directions, thereby improving the accuracy and stability of spectral acquisition.

[0045] When collecting spectra, the dual-axis displacement mechanism drives the portable Raman spectrometer 420 to move in a specified manner. The dual-axis displacement mechanism and the Raman spectrometer 420 cooperate with each other. Every time the Raman spectrometer 420 moves to a predetermined position, it is excited once, and Raman spectrum data of multiple points on the blade 60 are automatically collected.

[0046] In some embodiments, Figures 1 to 3 As shown, the blade bin 410 includes: a bin door 4110 and a bin body 4120. The bin door 4110 is rotatably disposed on the bin body 4120, and a blade cavity with an opening on one side is formed between the bin door 4110 and the bin body 4120.

[0047] In this embodiment, the bin body 4120 is the main part of the leaf bin 410, which forms a closed or semi-closed space for accommodating the leaves 60 of the target plant 50. The shape and size of the bin body 4120 are usually designed according to the size and shape of the target leaves 60 to ensure that the leaves 60 can remain stable therein. The material of the bin body 4120 needs to have a certain rigidity and stability to support the weight of the leaves 60 and keep the shape unchanged during the spectrum collection process.

[0048] The door 4110 is a part rotatably disposed on the chamber body 4120, and a blade cavity with an opening on one side is formed between the door 4110 and the chamber body 4120. This opening allows the blade 60 to easily enter and leave the blade cavity. The design of the door 4110 needs to ensure that it can tightly close the blade cavity to prevent external light or other interference factors from entering during the spectrum collection process. At the same time, the door 4110 also needs to have a certain degree of flexibility and durability so that it can maintain a good working condition during long-term use.

[0049] In this embodiment, the door 4110 is generally arranged below the chamber body 4120 and can be rotated downward to open the blade chamber. This design allows the blade 60 to conveniently fall into the blade chamber by gravity without additional operation. After the blade 60 enters the blade chamber, the door 4110 can be closed and locked to ensure that the blade 60 remains stable during the spectrum acquisition process.

[0050] In order to avoid interference caused by ambient light during Raman spectrum detection, the blade chamber 410 is preferably made of black material. Black material can absorb most of the light, thereby reducing the impact of reflection and scattering on the spectral signal. This design helps to improve the accuracy and stability of spectral data, making the identification of crop stress status more accurate and reliable.

[0051] In some embodiments, Figures 1 to 3 As shown, the blade bin 410 further includes: a blade clamp 4130 connected to the bin door 4110 for clamping the blade 60 disposed in the blade cavity.

[0052] Specifically, the clamping effect of the blade clamp 4130 can ensure that the blade 60 remains stable during the spectrum collection process and will not shake or move due to external factors (such as wind, vibration, etc.), which helps to improve the accuracy and stability of the spectral data.

[0053] In this embodiment, a plurality of blade clips 4130 are provided, and the plurality of blade clips 4130 are connected to the door 4110 at intervals. This design can ensure that the blade 60 can be well clamped at each position in the blade chamber. At the same time, the interval layout of the plurality of blade clips 4130 also helps to adapt to blades 60 of different sizes and shapes, so that the device can be applied to a wider range of crop types and blade 60 forms.

[0054] In order to further improve the adaptability and flexibility of the blade clamp 4130, the blade clamp 4130 in this embodiment is designed to be adjustable. Specifically, there are four adjustable blade clamps 4130 on the bottom surface of the blade bin 410 (i.e., the bin door 4110), and these blade clamps 4130 can be adjusted according to the size and shape of the blade 60 to ensure that the blade 60 can be firmly clamped and flattened. This design allows the device to be applicable to blades 60 of different sizes and shapes, and also ensures that the blade 60 is in the best state during the spectrum acquisition process.

[0055] After the blade 60 is fixed, the door 4110 needs to be closed to seal the blade chamber. In order to ensure that the blade 60 does not shake or move due to external factors during the spectrum acquisition process, the blade clamp 4130 needs to clamp the blade 60 in place and ensure that the blade 60 is within the focusing range of the Raman spectrometer 420.

[0056] In some embodiments, Figure 1 As shown, two support rods 100 extending in the lateral direction are formed on the fixed bracket 10; one end of the lateral displacement rod 20 is slidably mounted on one of the support rods 100, and the other end of the lateral displacement rod 20 is slidably mounted on the other support rod 100.

[0057] Specifically, a support rod 100 extending in the lateral direction is formed at the top of the fixed bracket 10, and a support rod 100 extending in the lateral direction is also formed at the bottom of the fixed bracket 10. The top end of the lateral displacement rod 20 can be slidably mounted on one of the support rods 100, and the bottom end of the lateral displacement rod 20 can be slidably mounted on the other support rod 100. By adjusting the position of the lateral displacement rod 20, lateral movement on the support rod 100 can be achieved, thereby adjusting the lateral position of the Raman spectrum acquisition assembly 40.

[0058] Accordingly, the longitudinal displacement platform 30 can be slidably mounted on the transverse displacement rod 20. The longitudinal displacement platform 30 is used to move along the longitudinal direction (i.e., the direction parallel to the transverse displacement rod 20) on the transverse displacement rod 20. By adjusting the position of the longitudinal displacement platform 30, the longitudinal position of the Raman spectrum collection component 40 can be further fine-tuned to ensure that it can accurately reach the designated position of the target plant 50.

[0059] Figure 4 FIG. 1 is a flow chart of a method for identifying the type of stress using a mobile crop stress on-body identification device provided by the present invention. Figure 4 As shown, the method includes the following: Step 401: Move the lateral displacement rod and the longitudinal displacement platform according to the spatial position of the target plant, so that the Raman spectrum collection component is located at the leaf position of the target plant.

[0060] Step 402: Fix the leaves of the target plant to the leaf bin, and use the dual-axis displacement mechanism to move the Raman spectrometer to collect Raman spectra at multiple sampling points on the leaves.

[0061] Step 403: pre-processing the collected Raman spectrum data, and selecting the spectral characteristic bands with the molecular information of the stress marker in the leaves.

[0062] Step 404: establishing an abiotic stress classification model based on the spectral characteristic bands with stress marker molecular information in the leaves to identify the stress type.

[0063] Specifically, when using the mobile crop stress on-body identification device of the embodiment of the present application, it is first necessary to accurately locate the target plant 50 according to its spatial position. This includes adjusting the positions of the lateral displacement rod 20 and the longitudinal displacement platform 30 so that the Raman spectrum acquisition component 40 can accurately reach the position of the leaf 60 of the target plant 50. The specific position of the leaf 60 of the target plant 50 is determined by naked eye observation or using other auxiliary equipment (such as a camera).

[0064] According to the lateral position of the target plant 50, the position of the lateral displacement rod 20 is slid on the support rod 100 to ensure that the lateral position of the Raman spectrum collection component 40 matches the target leaf 60. After the lateral displacement rod 20 is determined, the longitudinal position of the Raman spectrum collection component 40 is further fine-tuned by sliding the longitudinal displacement platform 30 on the lateral displacement rod 20 so that it can accurately reach the top of the target leaf 60.

[0065] Before the Raman spectrum is collected, the target leaf 60 needs to be fixed to the leaf bin 410 to prevent the accuracy of the spectral data from being affected by the shaking of the leaf 60 during the collection process. After the leaf 60 is fixed, the Raman spectrum collection assembly 40 is used to collect spectra at multiple sampling points on the leaf 60. These sampling points should be evenly distributed on the leaf 60 to ensure that the collected spectral data can fully reflect the stress state of the leaf 60.

[0066] The collected Raman spectral data are preprocessed, including noise removal, baseline correction, smoothing, etc., to improve the signal-to-noise ratio and accuracy of the spectral data. From the preprocessed spectral data, a spectral characteristic band having molecular information of stress markers in the leaf 60 is selected.

[0067] According to the selected spectral feature bands, an abiotic stress classification model is established using machine learning or deep learning algorithms. This model can automatically identify and classify the types of crop stress based on the input spectral feature band data. The preprocessed spectral feature band data is input into the established classification model, and the model automatically outputs the identification results of the stress type based on the input data. This result can help agricultural producers understand and respond to the stress status of crops in a timely manner, so as to take corresponding management measures to improve the yield and quality of crops.

[0068] In a specific embodiment, healthy wheat plants and wheat plants subjected to early low temperature stress are selected, and the fixed support 10 is placed next to the wheat to be tested. The lateral displacement rod 20 and the longitudinal displacement platform 30 are adjusted according to the positions of the target plant 50 and the target leaf 60, so that the Raman spectrum collection component 40 is at the position of the target leaf 60. The door 4110 at the bottom of the leaf bin 410 is opened, and the wheat leaf 60 is fixed and flattened using the four leaf clips 4130 on the door 4110. The door 4110 of the leaf bin 410 is closed, and the portable Raman spectrometer 420 is started, and the Raman spectrum is collected at multiple sampling points on the leaf 60 using a dual-axis displacement mechanism. Three sampling points are selected at a fixed distance from the tip to the root on both sides of the midrib of the leaf 60, and six Raman spectra are collected for each leaf 60. 25 healthy and low temperature treated wheat plants are selected as samples, and the second and third leaves of each wheat plant are selected to collect Raman spectrum data. After tuning, the excitation wavelength was selected as 785 nm, the laser power was 200 mW, and the integration time was 5 s.

[0069] The collected Raman spectral data were preprocessed, and baseline correction and smoothing were performed using airPLS and SG methods. The 600-1700 cm-1 part with 60 stress marker molecular information in leaves was selected. A classification model based on the random forest (RF) algorithm was established. After cross-validation for hyperparameter optimization, the model's accuracy in identifying low temperature stress in the test set of wheat reached 94.4%.

[0070] In another specific embodiment, Zhengdan 958 corn variety widely used in production and research is selected for cultivation, and low temperature and drought stress treatment is performed after normal growth to the three-leaf one-heart stage. Healthy, low temperature and drought corn plants are selected respectively, and the fixed bracket 10 is placed next to the corn to be tested. The lateral displacement rod 20 and the longitudinal displacement platform 30 are adjusted according to the position of the target plant 50 and the target leaf 60, so that the Raman spectrum acquisition system is at the target leaf 60 position. The door 4110 at the bottom of the leaf bin 410 is opened, and the corn leaf 60 is fixed and flattened using the four leaf clips 4130 on the door 4110. The door 4110 of the leaf bin 410 is closed, and the portable Raman spectrometer 420 is started, and the Raman spectrum is collected at multiple sampling points on the leaf 60 using a dual-axis displacement mechanism. Four sampling points are selected at a fixed distance from the tip to the root of the leaf on both sides of the midrib of the leaf 60, and 8 Raman spectra are collected for each leaf 60. Raman spectroscopy data were collected from 24 healthy corn leaves and 60 corn leaves under low temperature and drought stress. After optimization, the excitation wavelength was 785 nm, the laser power was 200 mW, and the integration time was 5 s.

[0071] The collected Raman spectral data were preprocessed, and baseline correction and smoothing were performed using airPLS and SG methods. The representative spectra and characteristic peaks after processing are shown in Figure 2. Figure 5 As shown. The 600-1700 cm-1 part with the molecular information of stress markers in leaves 60 was selected. A classification model based on the random forest (RF) algorithm was established. After cross-validation for hyperparameter optimization, the model had an accuracy rate of about 92.2% in identifying abiotic stress in the test set of corn. The results are shown in Figure 6 shown.

[0072] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A mobile crop stress on-body recognition device, characterized in that: include: A fixed bracket formed with a support rod extending in a transverse direction; A transverse displacement rod extending in the longitudinal direction and slidably disposed on the support rod for transverse movement on the support rod; A longitudinal displacement platform, slidably disposed on the transverse displacement rod, and configured to move longitudinally on the transverse displacement rod; The Raman spectrum acquisition component is arranged on the longitudinal displacement platform and is used to collect Raman spectra on the target plant, screen the spectral characteristic bands containing stress information, establish an abiotic stress classification model based on the spectral data of the band where the characteristic peak of the stress marker is located, and identify the stress type.

2. The mobile crop stress on-body recognition device according to claim 1, characterized in that: The Raman spectrum collection component includes: a blade chamber, a biaxial displacement mechanism and a Raman spectrometer; The leaf bin is arranged on the longitudinal displacement platform, and is formed with a leaf cavity, and the leaf cavity is used to accommodate leaves of the target plant; The detection end of the Raman spectrometer faces the blade cavity and is connected to the biaxial displacement mechanism, so that the position of the detection end of the Raman spectrometer facing the blade can be adjusted through the biaxial displacement mechanism.

3. The mobile crop stress on-body recognition device according to claim 2, characterized in that: The dual-axis displacement mechanism comprises: a movable platform, an X-axis moving component and a Y-axis moving component; The detection end of the Raman spectrometer is arranged on the movable platform, the movable platform is arranged on the X-axis moving component, and the X-axis moving component is used to drive the movable platform in a first lateral direction; the X-axis moving component is arranged on the Y-axis moving component, and the Y-axis moving component is used to drive the movable platform and the X-axis moving component in a second lateral direction.

4. The mobile crop stress on-body recognition device according to claim 2, characterized in that: The blade bin comprises: a bin door and a bin body, The bin door is rotatably arranged on the bin body, and the blade cavity with an opening on one side is formed between the bin door and the bin body.

5. The mobile crop stress on-body recognition device according to claim 2, characterized in that: The blade bin is made of black material.

6. The mobile crop stress on-body recognition device according to claim 4, characterized in that: The blade bin also includes: A blade clamp is connected to the compartment door and is used for clamping the blades arranged in the blade cavity.

7. The mobile crop stress on-body recognition device according to claim 6, characterized in that: A plurality of blade clamps are provided, and the plurality of blade clamps are connected to the bin door at intervals.

8. The mobile crop stress on-body recognition device according to claim 1, characterized in that: The fixing bracket is formed with two support rods extending in the transverse direction; One end of the lateral displacement rod can be slidably mounted on one of the support rods, and the other end of the lateral displacement rod can be slidably mounted on the other support rod.

9. The mobile crop stress on-body recognition device according to claim 8, characterized in that: The longitudinal displacement platform is slidably mounted on the transverse displacement rod.

10. A method for identifying the type of stress using the mobile crop stress on-body identification device according to any one of claims 1 to 9, characterized in that: include: According to the spatial position of the target plant, the lateral displacement rod and the longitudinal displacement platform are moved so that the Raman spectrum collection component is located at the leaf position of the target plant; The leaves of the target plant are fixed to the leaf chamber, and the Raman spectrometer is moved by a dual-axis displacement mechanism to collect Raman spectra at multiple sampling points on the leaves; Preprocess the collected Raman spectral data and select the spectral characteristic bands with molecular information of stress markers in leaves; An abiotic stress classification model was established based on the spectral characteristic bands containing molecular information of stress markers in leaves to identify the types of stress.

Images