All-terrain crop phenotype multi-source detection device

Abstract

The invention is suitable for the technical field of crop phenotype detection, and provides an all-terrain crop phenotype multi-source detection device which comprises an advancing system, a multi-source sensor quick connection module and a crop phenotype detection system. The device is suitable for automatic detection of crop phenotypes in a whole growth cycle in a field or a greenhouse, freely switches wheel-track modes according to different ground environments, and improves the comprehensive efficiency of crop phenotype detection. The sensor is convenient to replace and high in expansibility, based on an active light noise reduction and multi-source synchronous detection technology, physiological information and morphological characteristics of crops are accurately analyzed, and a multi-dimensional data system is constructed for crop phenotypic omics research, growth dynamic tracking and environmental factor regulation and control.

Description

Technical Field

[0001] The present invention belongs to the technical field of crop phenotype detection, and particularly relates to a multi-source detection device for all-terrain crop phenotypes. Background Art

[0002] Crop phenotype is a measurable characteristic jointly produced by genes and the environment, including comprehensive indicators such as plant morphology (such as plant height, canopy width, leaf area), physiological characteristics, and spectral response, and is the core basis for analyzing crop growth laws. High-quality detection and analysis technologies for crop phenotypes can provide key data support for variety improvement, gene function verification, research on environmental adaptation mechanisms, and intelligent agricultural management. Combining automated equipment to build an efficient and accurate crop phenotype detection system is becoming an important technical direction for promoting the development of agricultural digitization and intelligent breeding.

[0003] Existing crop phenotype detection equipment mainly obtains phenotype information by moving forward on wheels or transporting potted plants by belt. The wheels are prone to slipping when driving in the field, and the soil compaction is large. The belt type is only suitable for fixed-point detection of potted plants. The types of sensors carried by phenotype detection are few, resulting in poor analysis results. Moreover, the sensor switching depends on manual operation, and the coordination between mechanical automation control and the data detection module is insufficient, making it difficult to meet the compound requirements of high-throughput crop phenomics for precision and efficiency. Summary of the Invention

[0004] The purpose of the embodiments of the present invention is to provide a multi-source detection device for all-terrain crop phenotypes, aiming to solve the problems proposed in the above background art.

[0005] The embodiments of the present invention are implemented as follows. A multi-source detection device for all-terrain crop phenotypes includes a forward system, a multi-source sensor quick connection module, and a crop phenotype detection system; The forward system includes a frame. A crawler-type mobile module and a mobile wheel module are arranged on the frame. When the crawler-type mobile module moves during use, the mobile wheel module is in a contracted state and suspended; when the mobile wheel module moves during use, the mobile wheel module is in an expanded state and lifts the crawler-type mobile module. The multi-source sensor quick connection module is used for the automatic and quick replacement and docking of multi-source sensors; The crop phenotype detection system is used for collecting crop phenotype graphics and performing different information collections through the multi-source sensors in the multi-source sensor quick connection module.

[0006] A further technical solution is that the frame has a trapezoidal structure, and a square opening is provided at the front end of the frame and is connected with a detachable end cover.

[0007] Further technical solution: The crawler-type mobile modules are symmetrically installed on both sides of the frame. Specifically, it includes two triangular crawlers symmetrically installed on both sides of the frame. Inside the triangular crawlers, small tension wheels and large tension wheels are symmetrically distributed. Among them, there are three small tension wheels and two large tension wheels. The small tension wheels are arranged above the inner side of the triangular crawler, and the large tension wheels are arranged below the upper part of the inner side of the triangular crawler. Through the small tension wheels and large tension wheels, the triangular crawler forms an isosceles pentagon structure; each of the small tension wheels is installed on a second transmission shaft, and the second transmission shaft is connected to the output end of the brushless motor through a synchronous belt.

[0008] Further technical solution: The mobile wheel module includes a telescopic outer wheel and a servo motor. The telescopic outer wheel includes an outer hub, an X-shaped hinge structure, and an inner hub disk. The outer sides of the two arms of the X-shaped hinge structure are hinged to the outer hub, and the inner sides of the two arms of the X-shaped hinge structure are respectively hinged to the inner hub disk and a spiral sleeve; The output end of the servo motor is connected to the first transmission shaft through a two-stage reduction gearbox. The first transmission shaft is connected to a T-shaped lead screw, and an electromagnetic clutch is also arranged between the first transmission shaft and the T-shaped lead screw. The spiral sleeve is installed on the T-shaped lead screw; The spiral sleeve is also connected to the second transmission shaft through a sprocket chain mechanism.

[0009] Further technical solution: The arc-shaped outer surface of the outer hub is evenly covered with a styrene-butadiene rubber layer.

[0010] Further technical solution: The multi-source sensor quick-connection module includes a quick-connection module frame installed on the frame. A fixed plate is fixedly installed on the quick-connection module frame. A sliding plate is also slidably installed on the quick-connection module frame, and an electric telescopic rod for driving the sliding plate to slide is arranged on the quick-connection module frame. An L-shaped rod is hinged on both the fixed plate and the sliding plate. The free ends of the two L-shaped rods are both hinged to a lifting plate in the quick-connection module frame. Two quick-connection trays are placed on the lifting plate, and the top of the quick-connection tray is an upwardly protruding arc structure. Multi-source sensors are installed in each of the quick-connection trays; Above the quick-connection module frame, a double slide rail is also provided. A moving slider is slidably installed on the double slide rail. A spring lock mechanism is arranged at the bottom of the moving slider. The spring lock mechanism includes two rotating handles symmetrically installed in the moving slider. Each rotating handle is connected to the inner wall of the moving slider through a spring, and an electromagnet is correspondingly arranged at each rotating handle. The lower end of the rotating handle is a protruding arc structure.

[0011] Further technical solution: Two groups of crop phenotype detection systems are symmetrically arranged on the frame. The crop phenotype detection system includes a stainless-steel bracket installed on the frame. A gantry synchronous belt module with vertical freedom is installed on the stainless-steel bracket. The synchronous belt module is connected to a detection darkroom through a flange rod. The two detection darkrooms are interconnected through a square aluminum tube, and the double slide rails are installed in the square aluminum tube. An adjustable lamp group is installed in the detection darkroom. An annular guide rail is also installed inside the detection darkroom, and a connecting slider is installed on the annular guide rail. An RGB-D camera is installed on the connecting slider.

[0012] Further technical solution: The RGB-D camera is used to scan and obtain a point cloud with color information. The original data is lightened by using a voxel downsampling method, and then the noise is removed by point cloud filtering. Assuming the coordinates of the th point in the point cloud are , then the distance from point to any other point is : ; The average value of the distances between each point in the point cloud and any other point : ; The corresponding standard deviation : ; According to the determined multiple of the standard deviation , calculate whether the average distance from each point to other points in the neighborhood is within the truncation range , determine and remove the outlier points to achieve denoising. Use the ICP (Iterative Closest Point) algorithm for point cloud registration. Iteratively search for the maximum and minimum differences in the vertical spatial coordinates and the horizontal amplitude in the plant canopy point cloud, reconstruct the three-dimensional grid of the leaf point cloud and calculate the surface area integral, and finally obtain three-dimensional morphological parameters such as crop plant height, crown width, and leaf area.

[0013] Further technical solution: A black annular soft cloth is installed at the bottom of the detection darkroom through a fixed ring, and a crank-slider mechanism for driving the fixed ring to lift is also provided at the top of the detection darkroom. A stepping motor for driving the crank-slider mechanism is installed at the top of the detection darkroom, and the fixed ring is connected to the moving end of the crank-slider mechanism.

[0014] Further technical solution: The multi-source sensor includes a thermal imaging sensor and a hyperspectral camera, and the thermal imaging sensor and the hyperspectral camera are respectively installed in two quick-connect trays; The thermal imaging sensor is used to detect the thermal images of each leaf in the crop canopy, and combined with the visible light image for temperature calibration, image registration and denoising; Subsequently, by calculating the average temperature of the canopy and the daily temperature dynamic change curve, the key physiological parameters of crop water are inverted, and finally the phenotypic parameters of the crop are obtained; The hyperspectral camera is used to obtain continuous narrow-band spectral images. By extracting spectral reflectance, first derivative spectra, and spectral vegetation indices, and with the help of machine learning algorithms, a model of the quantitative relationship between spectral features and crop physiological characteristic parameters is constructed to invert crop phenotypic information.

[0015] A multi-source detection device for all-terrain crop phenotypes provided by an embodiment of the present invention is applicable to the automatic detection of crop phenotypes in the entire growth cycle in the field or greenhouse. It can freely switch between wheel and track modes for different ground environments, improving the comprehensive efficiency of crop phenotype detection. The quick connection module of the multi-source sensors carried by the device is automatically docked and locked with the crop phenotype detection system, and the replacement of the multi-source sensors is convenient and has strong expandability. The multi-source sensors carried by the device use double-track exchange sliding and circumferential scanning to obtain crop phenotype information in real time. Based on the active light noise reduction and multi-source synchronous detection technology, the physiological information and morphological characteristics of the crop are accurately analyzed, and a multi-dimensional data system is constructed for crop phenomics research, growth dynamic tracking, and environmental factor regulation. Description of the Drawings

[0016] Figure 1 It is a schematic structural diagram of a multi-source detection device for all-terrain crop phenotypes provided by an embodiment of the present invention; Figure 2 It is a schematic structural diagram of a retractable outer wheel in a multi-source detection device for all-terrain crop phenotypes provided by an embodiment of the present invention; Figure 3 It is a schematic structural diagram of a small tension wheel drive scheme in a multi-source detection device for all-terrain crop phenotypes provided by an embodiment of the present invention; Figure 4 It is a schematic structural diagram of a quick connection module for multi-source sensors in a multi-source detection device for all-terrain crop phenotypes provided by an embodiment of the present invention; Figure 5 It is a schematic structural diagram of a crop phenotype detection system in a multi-source detection device for all-terrain crop phenotypes provided by an embodiment of the present invention; Figure 6 It is a schematic internal structure diagram of a detection darkroom in a multi-source detection device for all-terrain crop phenotypes provided by an embodiment of the present invention; Figure 7 It is Figure 4 The enlarged view at A in

[0017] In the attached drawings: Forward system 1; Multi-source sensor quick-connect module 2; Crop phenotype detection system 3; Triangular crawler 4; Frame 5; Removable end cap 6; Small tension pulley 7; Large tension pulley 8; Telescopic outer wheel 9; Servo motor 10; Double-stage reduction gearbox 11; Cylindrical helical gear set 12; First transmission shaft 13; Electromagnetic clutch 14; T-shaped lead screw 15; First sprocket 16; Spiral sleeve 17; X-shaped hinge structure 18; Outer hub 19; Inner hub disk 20; Brushless motor 21; Synchronous belt 22; Second transmission shaft 23; Second sprocket 24; Quick-connect module frame 25; Fixed plate 26; L-shaped rod 27; Lifting plate 28; Multi-source sensor 29; Quick-connect tray 30; Slide plate 31; Electric telescopic rod 32; Double slide rail 33; Moving slider 34; Spring lock mechanism 35; Rotating handle 36; Electromagnet 37; Spring 38; Stainless steel bracket 39; Square aluminum tube 40; Detection darkroom 41; Black annular soft cloth 42; Crank-slider mechanism 43; Stepper motor 44; Flange rod 45; Synchronous belt module 46; Connecting slider 47; RGB-D camera 48; Annular guide rail 49; Adjustable lamp set 50. Detailed implementation manners

[0018] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the attached drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0019] The following describes in detail the specific implementation of the present invention with reference to specific embodiments.

[0020] As Figure 1 shown, a full-terrain crop phenotype multi-source detection device provided by an embodiment of the present invention includes a forward system 1, a multi-source sensor quick-connect module 2, and a crop phenotype detection system 3; The forward system 1 includes a frame 5. A crawler-type moving module and a moving wheel module are arranged on the frame 5. When the crawler-type moving module moves during use, the moving wheel module is in a contracted state and suspended; when the moving wheel module moves during use, the moving wheel module is in an expanded state and lifts the crawler-type moving module. The multi-source sensor quick-connect module 2 is used for the automatic and quick replacement and docking of the multi-source sensor 29. The crop phenotype detection system 3 is used for collecting crop phenotype graphics and performing different information collections through the multi-source sensor 29 in the multi-source sensor quick-connect module 2.

[0021] In an embodiment of the present invention, the forward system 1 has a freely switchable wheel-track dual mode. The track mode facilitates the stable forward movement of the detection vehicle in field environments such as rough and complex ground and large soil viscosity, and the wheel mode is suitable for fast driving on flat ground. During detection, the crop to be detected is covered by the crop phenotype detection system 3, and then the multi-source sensor 29 collects crop phenotype information.

[0022] As Figure 1 shown, as a preferred embodiment of the present invention, the frame 5 is integrally trapezoidal in structure and made of high-strength manganese steel. An electrical box for placing equipment such as a power supply and control system hardware is provided above it. A square opening is formed at the front end of the frame 5, and a detachable end cover 6 is connected by bolts, facilitating the installation and maintenance of the transmission components inside the frame 5.

[0023] As Figures 1 - 3 shown, as a preferred embodiment of the present invention, the tracked mobile module is symmetrically installed on both sides of the frame 5, specifically including two triangular tracks 4 symmetrically installed on both sides of the frame 5. Smaller tension wheels 7 and larger tension wheels 8 are symmetrically distributed inside the triangular track 4. Among them, three smaller tension wheels 7 are provided, and two larger tension wheels 8 are provided. The smaller tension wheels 7 are arranged above the inner side of the triangular track 4, and the larger tension wheels 8 are arranged below the inner side of the triangular track 4 above. The triangular track 4 forms an isosceles pentagon structure through the smaller tension wheels 7 and the larger tension wheels 8; each of the smaller tension wheels 7 is installed on a second transmission shaft 23, and the second transmission shaft 23 is connected to the output end of the brushless motor 21 through a synchronous belt 22.

[0024] In an embodiment of the present invention, the smaller tension wheels 7 at the same height are driving wheels, and its transmission method is: the brushless motor 21 transmits torque to the second transmission shaft 23 through the synchronous belt 22, the second transmission shaft 23 drives the smaller tension wheels 7 to rotate, and then drives the triangular track 4 to move forward. The independent drive makes the turning radius of the forward system 1 small and the flexibility high.

[0025] As Figures 1 - 3 shown, as a preferred embodiment of the present invention, the mobile wheel module includes a telescopic outer wheel 9 and a servo motor 10. The telescopic outer wheel 9 includes an outer hub 19, an X-shaped hinge structure 18, and an inner hub disc 20. The outer sides of the two arms of the X-shaped hinge structure 18 are hinged to the outer hub 19, and the inner sides of the two arms of the X-shaped hinge structure 18 are respectively hinged to the inner hub disc 20 and the spiral sleeve 17; The output end of the servo motor 10 is connected to the first transmission shaft 13 through a two-stage reduction gearbox 11. The first transmission shaft 13 is connected to a T-shaped lead screw 15, and an electromagnetic clutch 14 is also provided between the first transmission shaft 13 and the T-shaped lead screw 15. The spiral sleeve 17 is installed on the T-shaped lead screw 15; The spiral sleeve 17 is also connected to the second transmission shaft 23 through a sprocket and chain mechanism. Specifically, a second sprocket 24 is installed on the second transmission shaft 23, and a first sprocket 16 is installed on the spiral sleeve 17. The second sprocket 24 and the first sprocket 16 are connected by a chain, so as to transmit the torque of the second transmission shaft 23 to the spiral sleeve 17 through chain drive.

[0026] In the embodiment of the present invention, the transmission scheme of the T-shaped lead screw 15 is as follows: The power of the servo motor 10 is increased in torque by the cylindrical helical gear set 12 in the double-stage reduction gearbox 11 and then transmitted to the first transmission shaft 13. When the friction plates in the electromagnetic clutch 14 are closed, the first transmission shaft 13 drives the T-shaped lead screw 15 to rotate, pushing the spiral sleeve 17 to move, and driving the outer hub 19 to move radially along the spiral sleeve 17 through the X-shaped hinge structure 18 to control the wheel diameter expansion and contraction. When switching to the wheel drive mode, the friction plates in the electromagnetic clutch 14 are disconnected, and the spiral sleeve 17 is driven to rotate through the first sprocket 16, driving the retractable outer wheel 9 to rotate. At this time, the T-shaped lead screw 15 remains stationary with the spiral sleeve 17 due to self-locking, and the self-locking mechanism ensures the stability after the wheel diameter adjustment. The crawler and wheel drive modes share a set of power systems, reducing energy consumption and control complexity.

[0027] As a preferred embodiment of the present invention, the arc-shaped outer surface of the outer hub 19 is evenly covered with a wear-resistant styrene-butadiene rubber layer; As Figure 1 , Figure 4 and Figure 7 shown, as a preferred embodiment of the present invention, the multi-source sensor quick-connect module 2 includes a quick-connect module frame 25 installed on the frame 5. A fixing plate 26 is fixedly installed on the quick-connect module frame 25. A sliding plate 31 is also slidably installed on the quick-connect module frame 25, and an electric telescopic rod 32 for driving the sliding plate 31 to slide is provided on the quick-connect module frame 25. An L-shaped rod 27 is hinged to both the fixing plate 26 and the sliding plate 31. The free ends of the two L-shaped rods 27 are both hinged to a lifting plate 28 in the quick-connect module frame 25. Two quick-connect trays 30 are placed on the lifting plate 28, and the top of the quick-connect tray 30 is an upwardly protruding arc structure. A multi-source sensor 29 is installed in each quick-connect tray 30; Above the quick-connect module frame 25, a double slide rail 33 is also provided. A moving slider 34 is slidably installed on the double slide rail 33. A spring lock mechanism 35 is provided at the bottom of the moving slider 34. The spring lock mechanism 35 includes two rotating handles 36 symmetrically installed in the moving slider 34. Each rotating handle 36 is connected to the inner wall of the moving slider 34 through a spring 38, and an electromagnet 37 is correspondingly provided at each rotating handle 36. The lower end of the rotating handle 36 is a protruding arc structure.

[0028] In an embodiment of the present invention, the electric telescopic rod 32 pushes the skateboard 31 to slide outward, thereby driving the lifting plate 28 to run smoothly upward. The quick-connect tray 30 is docked with the spring lock mechanism 35 and interlocked. The arc structure at the lower end of the rotating handle 36 allows the arc structure at the top of the quick-connect tray 30 to be inserted. At this time, the spring is in a compressed state, realizing the interlocking function. When the electromagnet 37 is energized, the rotating handle 36 is attracted and releases the quick-connect tray 30, unlocking it. The moving slider 34 slides on the double slide rails 33 installed in parallel, thereby driving the multi-source sensor 29 in the quick-connect tray 30 to slide above the crop through the spring lock mechanism 35 for detection. Subsequently, it slides alternately to complete the double-region detection of the plant, realizing the real-time synchronous high-throughput detection of crop information by the multi-source sensor 29.

[0029] As Figure 1 , Figure 5 and Figure 6 shown, as a preferred embodiment of the present invention, two groups of crop phenotype detection systems 3 are symmetrically arranged on the frame 5. The crop phenotype detection system 3 includes a stainless-steel bracket 39 installed on the frame 5. A gantry synchronous belt module 46 with vertical freedom is installed on the stainless-steel bracket 39. The synchronous belt module 46 is connected to a detection darkroom 41 through a flange rod 45, and the detection darkroom 41 can be lowered to the ground to adapt to the crop phenotype detection with different growth spacings. The two detection darkrooms 41 are interconnected through a square aluminum tube 40, and the double slide rails 33 are installed in the square aluminum tube 40. An adjustable lamp group 50 is installed in the detection darkroom 41 to provide a stable light environment for phenotype detection; an annular guide rail 49 is also installed inside the detection darkroom 41, and a connecting slider 47 is installed on the annular guide rail 49. An RGB-D camera 48 is installed on the connecting slider 47.

[0030] In an embodiment of the present invention, a chute is opened at the bottom of the square aluminum tube 40 for the moving slider 34 on the double slide rails 33 to slide. By adjusting the position of the connecting slider 47 on the annular guide rail 49, the shooting angle of the RGB-D camera 48 can be freely adjusted. The RGB-D camera 48 is used to scan and obtain a point cloud with color information. By adopting the voxel downsampling method, the original data is lightweight, which is convenient for later analysis and calculation. Then, the point cloud is filtered to remove noise. Since the detection darkroom 41 of this device has a function of reducing light noise, most of the point cloud noise to be filtered is marginalized discrete points. Assuming that the coordinates of the th point in the point cloud are , then the distance from point to any other point is : ; The average value of the distances between each point in the point cloud and any other point : ; The corresponding standard deviation : ; According to the determined multiple of the standard deviation , calculate whether the average distance from each point to other points in the neighborhood is within the truncation range Judge and remove outliers to achieve denoising, effectively remove outliers in the point cloud, use the ICP (Iterative Closest Point) algorithm for point cloud registration, iteratively search for the maximum and minimum differences in the vertical spatial coordinates and the horizontal amplitude in the plant canopy point cloud, reconstruct the three-dimensional grid of the leaf point cloud and calculate the integral of its surface area, and finally obtain three-dimensional morphological parameters such as crop plant height, crown width, and leaf area.

[0031] As a preferred embodiment of the present invention, the detection darkroom 41 is made of aluminum alloy, and the cylindrical shape makes the internal light reflection more uniform. At the same time, the inner layer is coated with a barium sulfate coating with high diffuse reflection to improve the quality of crop phenotype imaging.

[0032] As Figure 5 shown, as a preferred embodiment of the present invention, the bottom of the detection darkroom 41 is installed with a black annular soft cloth 42 through a fixed ring, and the top of the detection darkroom 41 is also provided with a crank-slider mechanism 43 for driving the fixed ring to lift. A stepping motor 44 for driving the crank-slider mechanism 43 is installed on the top of the detection darkroom 41, and the fixed ring is connected to the moving end of the crank-slider mechanism 43.

[0033] In the embodiment of the present invention, during use, the stepping motor 44 drives the crank-slider mechanism 43 to move, thereby driving the fixed ring to lift. The black annular soft cloth 42 can avoid light leakage due to the gap between the detection darkroom 41 and the uneven ground, and actively reduce light noise to improve the quality of phenotype detection.

[0034] As a preferred embodiment of the present invention, the multi-source sensor 29 includes a thermal imaging sensor and a hyperspectral camera, and the thermal imaging sensor and the hyperspectral camera are respectively installed in two quick-connect trays 30; The thermal imaging sensor is used to detect the thermal images of each leaf in the crop canopy, perform temperature calibration, image registration, and denoising in combination with visible light images; then, by calculating the average canopy temperature and the daily temperature dynamic change curve, key physiological parameters such as crop water stress index and stomatal conductance are inverted, and finally phenotype parameters such as crop water status, photosynthetic efficiency, and canopy temperature distribution uniformity are obtained.

[0035] Hyperspectral cameras are used to obtain continuous narrow-band spectral images. Each pixel contains a rich spectral information curve. By extracting features such as spectral reflectance, first derivative spectra, and spectral vegetation indices, and using machine learning algorithms such as support vector machines (SVMs, Support Vector Machine) and random forests (RFs, Random Forest), a model of the quantitative relationship between spectral features and crop physiological characteristic parameters is constructed to invert key crop phenotype information such as chlorophyll content, saline-alkali stress degree, and biomass.

[0036] As a preferred embodiment of the present invention, it further includes a control system. The control system uses the STM32F407 microcontroller as the core processing unit to achieve coordinated control of multi-threaded actuators through integrated design. The main control unit establishes communication connections with each motor, multi-source sensor, trigger relay, etc. through the CAN (Controller Area Network) bus protocol, and at the same time receives the detection data of the multi-source sensors in real time. The system uses WiFi / 4G dual-mode wireless transmission technology to achieve two-way communication with the remote terminal. The upper computer performs fusion analysis on multi-source heterogeneous data through deep learning algorithms and dynamically presents the processing results on the visualization operation interface. At the same time, this design supports cross-platform interaction between the mobile terminal and the PC terminal. Operators can adjust the device parameters in real time through the cloud management platform and obtain multi-dimensional evaluation reports on the crop growth trend, which is beneficial to constructing a complete intelligent agricultural Internet of Things control system.

[0037] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An all-terrain crop phenotype multi-source detection device, characterized in that, It includes an advancing system, a multi-source sensor quick-connect module, and a crop phenotype detection system; The advancing system includes a frame. A crawler-type moving module and a moving wheel module are arranged on the frame. When the crawler-type moving module moves during use, the moving wheel module is in a retracted state and suspended; when the moving wheel module moves during use, the moving wheel module is in an expanded state and lifts the crawler-type moving module. The multi-source sensor quick-connect module is used for the automatic and quick replacement and docking of multi-source sensors; The crop phenotype detection system is used for collecting crop phenotype graphics and performing different information collections through the multi-source sensors in the multi-source sensor quick-connect module.

2. The all-terrain crop phenotype multi-source detection device according to claim 1, characterized in that, The frame is in a trapezoidal structure. A square opening is formed at the front end of the frame, and a detachable end cover is connected.

3. The all-terrain crop phenotype multi-source detection device according to claim 1, characterized in that The crawler-type moving module is symmetrically installed on both sides of the frame. Specifically, it includes two triangular crawlers symmetrically installed on both sides of the frame. Small tension wheels and large tension wheels are symmetrically distributed inside the triangular crawlers. Among them, three small tension wheels are provided, and two large tension wheels are provided. The small tension wheels are arranged above the inner side of the triangular crawler, and the large tension wheels are arranged below the upper part of the inner side of the triangular crawler. The triangular crawler forms an isosceles pentagon structure through the small tension wheels and large tension wheels; each of the small tension wheels is installed on a second transmission shaft, and the second transmission shaft is connected to the output end of the brushless motor through a synchronous belt.

4. The all-terrain crop phenotype multi-source detection device according to claim 3, wherein The moving wheel module includes a telescopic outer wheel and a servo motor. The telescopic outer wheel includes an outer hub, an X-shaped hinge structure, and an inner hub disc. The outer sides of the two arms of the X-shaped hinge structure are hinged to the outer hub, and the inner sides of the two arms of the X-shaped hinge structure are respectively hinged to the inner hub disc and a spiral sleeve; The output end of the servo motor is connected to a first transmission shaft through a two-stage reduction gearbox. The first transmission shaft is connected with a T-shaped lead screw, and an electromagnetic clutch is also arranged between the first transmission shaft and the T-shaped lead screw. The spiral sleeve is installed on the T-shaped lead screw; The spiral sleeve is also connected to the second transmission shaft through a sprocket chain mechanism.

5. The all-terrain crop phenotype multi-source detection device according to claim 4, characterized in that, The arc-shaped outer surface of the outer hub is evenly covered with a styrene-butadiene rubber layer.

6. The all-terrain crop phenotype multi-source detection device according to claim 1, characterized in that The multi-source sensor quick-connect module includes a quick-connect module frame installed on the frame. A fixed plate is fixedly installed on the quick-connect module frame. A slide plate is also slidably installed on the quick-connect module frame, and an electric telescopic rod for driving the slide plate to slide is arranged on the quick-connect module frame. An L-shaped rod is hinged on both the fixed plate and the slide plate. The free ends of the two L-shaped rods are both hinged to a lifting plate in the quick-connect module frame. Two quick-connect trays are placed on the lifting plate, and the top of the quick-connect tray is an upwardly protruding arc structure. Multi-source sensors are installed in each of the quick-connect trays; A double slide rail is also arranged above the quick-connect module frame. A moving slider is slidably installed on the double slide rail. A spring lock mechanism is arranged at the bottom of the moving slider. The spring lock mechanism includes two rotating handles symmetrically installed in the moving slider. Each of the rotating handles is connected to the inner wall of the moving slider through a spring, and an electromagnet is correspondingly arranged at the rotating handle. The lower end of the rotating handle is a protruding arc structure.

7. The all-terrain crop phenotype multi-source detection device according to claim 6, wherein There are two sets of crop phenotype detection systems symmetrically arranged on the frame. The crop phenotype detection system includes a stainless-steel bracket installed on the frame. A gantry synchronous belt module with vertical direction freedom is installed on the stainless-steel bracket. The synchronous belt module is connected to a detection darkroom through a flange rod. The two detection darkrooms are interconnected through a square aluminum tube, and the double slide rail is installed in the square aluminum tube. An adjustable lamp group is installed in the detection darkroom. An annular guide rail is also installed inside the detection darkroom, and a connecting slider is installed on the annular guide rail. An RGB-D camera is installed on the connecting slider.

8. The all-terrain crop phenotype multi-source detection device according to claim 7, characterized in that, The RGB-D camera is used to scan and obtain a point cloud with color information. The original data is lightened by using the voxel downsampling method, and then the noise is removed by point cloud filtering. Suppose the coordinate of the -th point in the point cloud is , then the distance from point to any other point is : ； Average distance between each point in the point cloud and any other point : ； The corresponding standard deviation : ； According to the multiple of the determined standard deviation , calculate whether the average distance from each point to other points in the neighborhood is within the truncation range , determine and remove outliers to achieve denoising, use the ICP algorithm for point cloud registration, perform extreme value iteration search for the difference between the maximum and minimum values of the vertical spatial coordinates and the horizontal amplitude in the plant canopy point cloud, reconstruct the three-dimensional grid of the leaf point cloud and calculate the integral of its surface area, and finally obtain three-dimensional morphological parameters such as crop plant height, crown width and leaf area.

9. The all-terrain crop phenotype multi-source detection device according to claim 7, wherein, A black annular soft cloth is installed at the bottom of the detection darkroom through a fixed ring, and a crank-slider mechanism for driving the fixed ring to lift is also provided at the top of the detection darkroom. A stepping motor for driving the crank-slider mechanism is installed at the top of the detection darkroom, and the fixed ring is connected to the mobile end of the crank-slider mechanism.

10. The all-terrain crop phenotype multi-source detection device according to claim 6, wherein The multi-source sensor includes a thermal imaging sensor and a hyperspectral camera, and the thermal imaging sensor and the hyperspectral camera are respectively installed in two quick-connect trays; The thermal imaging sensor is used to detect the thermal images of each leaf in the crop canopy, perform temperature calibration, image registration and denoising in combination with visible light images; then, by calculating the average temperature of the canopy and the daily temperature dynamic change curve, the key physiological parameters of crop water are inverted, and finally the phenotype parameters of the crop are obtained; The hyperspectral camera is used to obtain continuous narrow-band spectral images. By extracting spectral reflectance, first derivative spectra, and spectral vegetation indices, and with the help of machine learning algorithms, a model of the quantitative relationship between spectral features and crop physiological characteristic parameters is constructed to invert crop phenotype information.

Images