Adjustable greenhouse crop phenotype image acquisition and environment monitoring control device

Abstract

The utility model discloses an adjustable greenhouse crop phenotype image acquisition and environment monitoring control device, and mainly solves the problems that an existing image acquisition device needs manual measurement, the manual measurement efficiency is low, and the existing phenotype acquisition device is difficult to quickly and comprehensively acquire three-dimensional phenotype data of crops. The device comprises an image acquisition module, a detection processing module, an environmental parameter acquisition module, a motion control module and an actuator module. The image acquisition module acquires a three-dimensional image of a plant and transmits the acquired three-dimensional image of the plant to the detection processing module, and the detection processing module calculates a current developmental period of the plant based on the three-dimensional image of the plant and sets an environment parameter range of plant growth according to the current developmental period. According to the utility model, unsafe factors of the current environment of crops can be accurately and rapidly monitored and pre-warned, so that the crops grow healthily.

Description

TECHNICAL FIELD

[0001] The utility model belongs to the technical field of wisdom agricultural greenhouse, concretely, it relates to an adjustable greenhouse crop phenotype image acquisition and environment monitoring control device. BACKGROUND

[0002] China is a big agricultural country, and needs to cultivate better new crop varieties to ensure China's food security. At present, the digital acquisition of crop phenotype is an important basis for studying crop growth and development, and cultivating better crop quality also needs to control various environmental factors. By using phenotype acquisition and analysis technology, combined with genomic information, the environment can be monitored to cultivate better new crops with higher yield, more resistance and more stress tolerance.

[0003] Due to the limited space in the greenhouse environment, it is difficult for staff to use unmanned aerial vehicles, wisdom agricultural machines and other common equipment to obtain plant phenotypes, and the use of handheld devices to collect plant phenotypes is inefficient, and the crop phenotype acquisition method is generally contact type acquisition, which inevitably damages the plants and affects the normal growth of crops.

[0004] The light intensity, air temperature and humidity, and soil temperature and humidity in the greenhouse are crucial to crop growth, and adverse weather conditions such as cloudy days, rain and snow can have a huge impact on agricultural production, directly affecting the degree of crop growth.

[0005] Modern greenhouses often have light intensity, air temperature and humidity, and soil temperature and humidity sensors to monitor and warn the light, temperature, and water environment in the greenhouse to ensure the normal growth of crops in the greenhouse. UTILITY MODEL CONTENT

[0006] The utility model aims at providing an adjustable greenhouse crop phenotype image acquisition and environment monitoring control device, mainly solving the problems of the existing image acquisition device needing manual measurement, low labor measurement efficiency, and the existing phenotype acquisition device being difficult to quickly and comprehensively obtain three-dimensional phenotype data of crops.

[0007] To achieve the above-mentioned purpose, the utility model adopts the following technical solutions:

[0008] An adjustable greenhouse crop phenotype image acquisition and environment monitoring control device, comprising an image acquisition module, a detection processing module, an environment parameter acquisition module, a motion control module, and an actuator module; the image acquisition module acquires a three-dimensional image of a plant and transmits the acquired three-dimensional image of the plant to the detection processing module, the detection processing module calculates the current growth stage of the plant based on the three-dimensional image of the plant and obtains phenotype parameters, and sets the environment parameter range for plant growth according to the current growth stage;

[0009] The detection processing module comprises a communication unit, a processing unit, an alarm unit, a server and a battery; an output end of the environment parameter acquisition module is electrically connected with the communication unit, an output end of the image acquisition module is electrically connected with the processing unit, the communication unit is electrically connected with the processing unit, the communication unit is connected with the server and the alarm unit through a WIFI wireless network respectively; and the processing unit sends data to the server for use after being powered by the battery.

[0010] Further, in the utility model, the image acquisition module includes a camera, the model of the camera is AzureKinect depth camera, and the camera is electrically connected with the processing unit.

[0011] Further, in the utility model, the environment parameter acquisition module includes an illumination sensor, an air temperature and humidity sensor and a soil temperature and humidity sensor.

[0012] Further, in the utility model, the executor module includes a relay electrically connected with the processing unit, and a roller shutter motor, a light supplement lamp, a film rolling motor and an electromagnetic water valve connected with the relay.

[0013] Further, the utility model also includes a mounting support arranged in the greenhouse, the mounting support includes four support columns for supporting, a crossbeam arranged on the support columns, two rack rails arranged on the two ends of the crossbeam, left and right moving sliders arranged on the rack rails and engaged with the rack rails through drive motors, front and rear moving sliders arranged on the rack rails and engaged with the rack rails through drive motors, a rack bar fixedly connected with the left and right moving sliders at two ends and engaged with the two rack rails, a motor arranged in the left and right moving sliders for driving the rack bar, and a motor arranged in the front and rear moving sliders for driving the camera; wherein the side wall of the left and right moving slider is slidably connected with the side wall of the rack rail, the drive motor is fixedly installed in the left and right moving slider, the front and rear moving sliders are slidably connected with the rack bar, and the drive motor is fixedly installed in the front and rear moving sliders; the illumination intensity sensor, the air temperature and humidity sensor and the soil temperature and humidity sensor are arranged on the support columns through mounting rings.

[0014] Further, in the utility model, the motor is controlled through a motion control module, the motion control module comprises a motor driver and a power storage unit; the power storage unit supplies power to the motor driver, the motor driver is electrically connected with the processing unit, and the motor driver is fixedly installed below the front and rear moving sliders.

[0015] Further, in the utility model, the support column adopts a support column with a telescopic rod; the upper end of the telescopic rod is connected with the crossbeam.

[0016] Further, in the utility model, the communication unit contains the wireless transmission WIFI module, the processing unit contains the raspberry.

[0017] Compared with the prior art, the utility model has the following beneficial effects:

[0018] The greenhouse adjustable crop phenotype image acquisition and environment monitoring control device has high accuracy, does not affect normal growth of crops, can automatically adjust environmental parameters according to crop growth periods, accurately and quickly monitors and early warns unsafe factors of the current environment of crops, and makes the crops grow healthily. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 It is the principle block diagram of the utility model.

[0020] Figure 2 It is the bottom structure schematic view of the upper part of the mounting bracket in the embodiment of the utility model.

[0021] Figure 3 It is the whole structure schematic view of the mounting bracket in the embodiment of the utility model.

[0022] Figure 4 It is the motor structure schematic view in the left and right movement sliding block in the embodiment of the utility model.

[0023] Figure 5 It is the motor structure schematic view in the front and rear movement sliding block in the embodiment of the utility model.

[0024] Wherein, the name corresponding to the reference sign is:

[0025] 1-image acquisition module, 2-detection processing module, 3-environmental parameter acquisition module, 4-mounting bracket, 11-camera, 21-communication unit, 22-processing unit, 23-alarm unit, 24-server, 25-battery, 31-illumination intensity sensor, 32-air temperature and humidity sensor, 33-soil temperature and humidity sensor, 41-support column, 42-cross beam, 43-toothed rail, 44-left and right movement sliding block, 45-rack, 46-motor, 47-mounting ring, 48-motor driver, 49-telescopic rod, 410-front and rear movement sliding block, 411-gear. DETAILED DESCRIPTION

[0026] The utility model will be further described below in combination with the drawings and embodiments, and the mode of the utility model includes but is not limited to the following embodiments.

[0027] As Figures 1 to 3The utility model discloses a kind of adjustable greenhouse crop phenotypic image acquisition and environmental monitoring control device, including image acquisition module 1, detection processing module 2, environmental parameter acquisition module 3, motion control module and actuator module;The image acquisition module 1 acquires plant three-dimensional image, and the plant three-dimensional image collected is transmitted to the detection processing module 2, the detection processing module 2 is based on plant three-dimensional image calculation plant current development period, according to the current development period needs to set the environmental parameter range of plant growth. Among them, the environmental parameter acquisition module 3 includes illumination intensity sensor 31, air temperature and humidity sensor 32 and soil temperature and humidity sensor 33. The illumination intensity sensor 31 is used to collect the illumination intensity in greenhouse, the air temperature and humidity sensor 32 is used to collect the air temperature and humidity in greenhouse, and the soil temperature and humidity sensor 33 is used to collect the temperature and humidity of soil in greenhouse. The environmental parameter acquisition module 3 is transmitted to communication unit 21 after the environmental data collected by wireless WIFI module transmission to server for saving and viewing real-time data.

[0028] In the embodiment, the plant current development period based on plant three-dimensional image calculation is a kind of prior art, and the article "plant growth stage recognition method based on improved ResNet18 and the implementation of intelligent plant light supplementing" discloses that after the image acquisition module is completed to image acquisition, the existing image processing technology can be used to realize the processing of image, to realize the plant current development period based on plant three-dimensional image calculation. Among them, the specific process of image processing is: after shooting, the image is transmitted to processing unit, and the processing unit is preprocessed to image data, then the processed image data is transmitted to communication unit, and the communication unit is transmitted to server by wireless WIFI module to do further processing. Server is handled and feature extraction to image by installing Python image processing library, machine learning framework, server is first to RGB and depth image is preprocessed, such as denoising, then RGB image is spatial scale transformation, normalization and standardization processing, and depth image is normalized. Use deep learning model to identify the growth stage of plant, map each pixel in depth image to three-dimensional point in camera coordinate system, to generate three-dimensional point cloud data, then extract plant image key features.

[0029] In the embodiment, the detection processing module 2 comprises a communication unit 21, a processing unit 22, an alarm unit 23, a server 24 and a battery 25; the output end of the environment parameter acquisition module 3 is electrically connected with the communication unit 21, the output end of the image acquisition module 1 is electrically connected with the processing unit 22, the communication unit 21 is electrically connected with the processing unit 22, the communication unit 21 is connected with the server 24 and the alarm unit 23 through a wireless network respectively; the processing unit 22 sends data to the server 24 for use after power supply. The communication unit 21 comprises a wireless transmission WIFI module, and the processing unit 22 comprises a Raspberry Pi. In other embodiments, the communication unit 21 comprises an extended Ethernet interface. Through the Ethernet interface, the environment parameter information and the crop phenotype parameter can be sent or acquired by using a network, and the intelligent degree and the safety warning function are improved.

[0030] In the embodiment, the device further comprises a mounting bracket 4 arranged in the greenhouse; the greenhouse is humid, and the mounting bracket 6 uses an aluminum alloy frame to prevent rust. The mounting bracket 4 comprises four support columns 41 for support, a cross beam 42 arranged on the support columns 41, two toothed rails 43 arranged on both ends of the cross beam 42, a rack 45 fixedly connected with the left and right moving sliders 44 at both ends and meshingly connected with the two toothed rails 43, a motor 46 (corresponding to the motor 1 and the motor 2 in FIG. 4, corresponding to the motor structure schematic diagram in FIG. 5) arranged in the left and right moving sliders 44 for driving the rack 45, and a front and rear moving slider 410 arranged on the rack 45 and meshingly connected with the rack 45 through a gear 411 of the driving motor 46; wherein the front and rear moving slider is in sliding connection with the rack, the side wall of the left and right moving slider 44 is in sliding connection with the side wall of the toothed rail 43, and the driving motor 46 (corresponding to the motor 3 in FIG. 4, corresponding to the motor structure schematic diagram in FIG. 5) is arranged in the front and rear moving slider 410. Figure 1 Figure 4 Figure 1 Figure 5 The light sensor 31, the air temperature and humidity sensor 32 and the soil temperature and humidity sensor 33 are arranged on the support columns 41 through a mounting ring 47. The motor 46 is controlled through a motion control module, and the motion control module comprises a motor driver 48 and a power storage unit; the power storage unit supplies power to the motor driver 48, the motor driver 48 is connected with the processing unit 22, and the motor driver 48 is fixedly arranged below the front and rear moving slider 410.

[0031] ​​​The motor driver 48 in this embodiment adopts a stepper motor driver model DRV8825, which supports micro-step control. Each stepper motor driver 48 controls one stepper motor, and the output end is directly connected to the coil of the stepper motor. The stepper motor model is NEMA 17. The motor driver 48 is connected to the processing unit 22 (Raspberry Pi) through the GPIO pin. The GPIO pin of the processing unit 22 (Raspberry Pi) is connected to the STEP and DIR pins of the motor driver 48. The processing unit 22 (Raspberry Pi) controls the speed and direction of the motor through the PWM signal. The STEP pin receives the PWM signal to control the motor speed, and the DIR pin controls the motor direction. The processing unit 22 contains a Raspberry Pi, which is a prior art technology with widely used and disclosed functions and characteristics (see Raspberry Pi official website [https: / / www.raspberrypi.org]). The model of the Raspberry Pi in the processing unit 22 is Raspberry Pi 4 Model B, which has strong computing power and is suitable for processing complex image processing tasks. Referring to Figure 2 , the Raspberry Pi in the processing unit 22 is based on the Raspberry Pi OS operating system. The GPIO pin of the Raspberry Pi is connected to the direction control pin (DIR) and the step pin (STEP) of the three motor drivers 48. The Raspberry Pi controls the rotation direction and step action of the motor by outputting high and low level signals. The communication unit 21, the processing unit 22, the motor driver 48, and the camera 11 are located below the front and rear moving slider 410 and are fixedly connected to the front and rear moving slider 410. The side wall of the left and right moving slider 44 is slidingly connected to the side wall of the tooth rail 43, and the front and rear moving slider 410 is slidingly connected to the tooth bar 45. The motor 46 is fixedly connected to the front and rear moving slider 410 and the left and right moving slider 44. The gear 411 is directly fixedly connected to the motor 46, and the left and right moving slider 44 and the front and rear moving slider 410 are fixedly connected to the motor 46. The rotation of the motor 46 drives the rotation of the gear 411 in multiple degrees of freedom. When the gear 411 rotates, the teeth of the gear 411 mesh with the teeth of the tooth bar 45 and the tooth rail 43, so that the gear 411 moves linearly along the tooth bar and the tooth rail. The gear 411 is fixedly connected to the motor 46, and the motor 46 moves linearly along the tooth bar 45 and the tooth rail 43 by being fixedly connected to the gear 411. The motor 46 is fixedly connected to the left and right moving slider 44 and the front and rear moving slider 410, thereby pushing the front and rear moving slider 410 to move on the tooth bar 45 and pushing the left and right moving slider 44 to move on the tooth rail 43. The camera 11 is fixedly connected below the front and rear moving slider 410, and the front and rear moving slider 410 and the left and right moving slider 44 move to drive the camera 11 to move, thereby shooting three-dimensional images of plants from multiple directions.

[0032] The camera is connected with the Raspberry Pi through the USB interface, and the photographed photo data is transmitted to the Raspberry Pi in real time. The Raspberry Pi uses PNG lossless compression RGB and depth image, and the compression level is set to 6 and the quality parameter is set to 95. The Raspberry Pi automatically triggers the camera to take pictures every 2 hours, and uploads the photo data to the server through the communication unit. After the system starts, the Raspberry Pi initializes the motor driver 48, sets the working mode of the GPIO pin, and sends an initialization signal. The Raspberry Pi obtains the control instruction by polling the server, and executes the corresponding motion control or basic motion instruction. The camera uses an Azure Kinect depth camera, and the shooting angle is 90 degrees perpendicular to the ground, which can capture three-dimensional phenotype data of crops. The camera is connected with the processing unit through the USB interface, and automatically takes RGB and depth images of crops every 2 hours. Taking tomatoes as the experimental object, the improved ResNet18 is used to identify the seedling stage, flowering and fruit setting stage and fruit ripening stage of tomatoes, and the leaf area index, plant height, leaf number and shape and other key growth characteristics of the plant are obtained after the depth image is converted into point cloud data. After shooting, the photographed image is transmitted to the processing unit, the processing unit compresses the image data, and then the processed image data is transmitted to the communication unit, and the communication unit transmits the image data to the server through the wireless WIFI module for further processing.

[0033] The server calls the Pytorch algorithm tool library to perform spatial scale transformation, normalization and standardization processing on the RGB image, and then normalizes the depth image. The improved ResNet18 model is used to identify the seedling stage, flowering and fruit setting stage and fruit ripening stage of tomatoes, the depth separable convolution is used instead of the traditional convolution, and the SE module (Squeeze and excitation module) is introduced to improve the efficiency and accuracy of model task processing. The improved model is trained combined with the early stopping method and the learning rate decay mechanism to avoid overfitting. In the seedling stage identification, the model focuses on the leaf and root features; for the flowering and fruit setting stage, the model focuses on the flower part; in the fruit ripening stage, the model focuses on the fruit part. After the depth image is converted into point cloud data, each pixel corresponds to a three-dimensional point (x, y, z), and the plant phenotype data can be extracted by analyzing the point cloud data. The leaf area index is calculated through the point cloud data, the plant height is determined through the maximum Z value in the point cloud data, and the number and shape of the leaves can be counted by analyzing the leaf outline of the point cloud data. The manager can master the plant growth status in time by means of the predicted growth stage result and the obtained phenotype data of the server, and take corresponding measures to change the environmental parameters of the greenhouse according to the current plant development stage, thereby improving the intelligent degree of the greenhouse.

[0034] After the system starts working, the Raspberry Pi of the processing unit initializes the motor driver 48, judges whether the server has information, executes the control motion instruction if there is an instruction, executes the basic motion instruction if there is no instruction, and realizes the stop, forward and backward movement and left and right movement of the monitoring mechanism. After shooting, the shot image is transmitted to the server to wait for further processing.

[0035] In the embodiment, the alarm unit adopts a warning light which is arranged on the support through a mounting ring. When the environmental parameter exceeds the threshold of the suitable environmental parameter, the alarm unit will warn the unsafe environmental parameter information. The image acquisition module 1 adopts an Azure kinect depth camera, and adopts a WFOV (Wide Field of View) mode during shooting, and the pixel is 1024x1024. The depth imaging principle of the Azure kinect depth camera is the TOF (Time of Flight) principle, the flight time of the light signal is measured through sending and receiving light pulses, so as to calculate the distance between the object and the camera. The preset time is 2 hours. When setting the camera, it should be ensured that the phenotype of the plant in the image shot by the camera is clear.

[0036] In the embodiment, the support 41 adopts a support with a telescopic rod 49; the telescopic rod 49 is connected with the cross beam 42 with a sliding rail above and the support 41 below. The camera 11 can run above the crops through the rack 43 and the gear 45, which avoids the trouble caused by repeatedly transferring the measuring equipment by the measurer, and the telescopic rod can be adjusted according to the height of the crops, which avoids that the collected plant image information is incomplete due to the too high height of the crops, or the collected crop image information is not clear due to the too low height of the crops.

[0037] The server is deployed with a web server for providing a remote access interface and a MySQL database for storing environmental parameter data, image data, and operation logs. The management personnel can access the web interface through a browser to view real-time data charts and understand the current environmental status at any time, and query historical data in the MySQL database. Based on the crop light intensity, air temperature and humidity, and soil temperature and humidity, and other environmental parameter information obtained or converted from the environmental parameter acquisition module, the server determines whether it is within the safety threshold range. If it exceeds the range, the server can send alarm information to the alarm unit and the processing unit through the WIFI interface of the communication unit according to the preset environmental parameter threshold. The alarm unit will alert the environmental parameter information that needs to be alarmed, such as light, air temperature and humidity, and soil temperature and humidity. The processing unit will control the actuator module to automatically control the actuator module to adjust. The Raspberry Pi outputs control signals through the GPIO pin. The relay uses a 5V 4-way relay module, each way of the relay includes an input signal end, a VCC pin, a GND pin, and an output end. The battery provides DC power for the relay module and the Raspberry Pi. The external power supply provides AC 220V AC power for the motor and the load (roller shutter motor, light supplement lamp, film rolling motor, electromagnetic water valve).

[0038] The input signal end of the relay module is connected to the GPIO pin of the Raspberry Pi, the VCC pin of the relay module is connected to the 5V pin of the Raspberry Pi, and the GND pin of the relay module is connected to the GND pin of the Raspberry Pi.

[0039] The output end of the relay module is connected to the power line of the load (roller shutter motor, light supplement lamp, film rolling motor, electromagnetic water valve). The power line of the load is connected to the external AC 220V AC power supply. The positive electrode of the battery 25 is connected to the 5V pin of the Raspberry Pi and the VCC pin of the relay module. The negative electrode of the battery 25 is connected to the GND pin of the Raspberry Pi and the GND pin of the relay module.

[0040] The Raspberry Pi outputs high and low level signals through the GPIO pin to control the on-off state of the relay module. When the Raspberry Pi outputs a low level, the input signal end of the relay module is turned on, the output end is closed, and the load is powered on to run. When the Raspberry Pi outputs a high level, the input signal end of the relay module is disconnected, and the load is powered off to stop running.

[0041] If the greenhouse light intensity is lower than the threshold, the light supplement lamp is automatically turned on. If the greenhouse temperature is higher than the threshold, the film rolling motor is automatically turned on to open the film for ventilation. If the greenhouse temperature is lower than the threshold, the roller shutter motor is automatically turned on to open the roller shutter for heat preservation. If the soil humidity is lower than the threshold, the electromagnetic water valve is automatically turned on to release water. The staff can also use the server to send control instructions, and the processing unit controls the actuator module to ventilate, control temperature, and water.

[0042] The above embodiment is only one of the preferred embodiments of the present application, and should not be used to limit the protection scope of the present application, but any insignificant modification or polishing made in the main design idea and spirit of the present application, the technical problems solved are still consistent with the present application, and should be included in the protection scope of the present application.

Claims

1. An adjustable greenhouse crop phenotype image acquisition and environmental monitoring control device, characterized in that, The device comprises an image acquisition module (1), a detection processing module (2), an environmental parameter acquisition module (3), a motion control module and an actuator module; the image acquisition module (1) acquires plant images and transmits the acquired plant images to the detection processing module (2); the detection processing module (2) calculates the current growth stage of the plant based on the plant images and obtains phenotype data, and sets the environmental parameter range for plant growth according to the current growth stage; The detection processing module (2) comprises a communication unit (21), a processing unit (22), an alarm unit (23), a server (24) and a battery (25); the output end of the environmental parameter acquisition module (3) is electrically connected with the communication unit (21); the output end of the image acquisition module (1) is electrically connected with the processing unit (22); the communication unit (21) is electrically connected with the processing unit (22); the communication unit (21) is connected with the server (24) and the alarm unit (23) through WIFI wireless network respectively; the processing unit (22) sends data to the server (24) for use after being powered by the battery (25); The actuator module comprises a relay electrically connected with the processing unit (22), a roller shutter motor, a light supplement lamp, a film rolling motor and an electromagnetic water valve connected with the relay; The device further comprises a mounting bracket (4) arranged in the greenhouse; the mounting bracket (4) comprises four support columns (41) for support, a cross beam (42) arranged on the support columns (41), two toothed rails (43) arranged on the cross beam (42), a rack (45) having two ends fixedly connected with left and right moving sliders (44) and meshingly connected with the two toothed rails (43), a motor (46) arranged in the left and right moving sliders (44) for driving the rack (45), and front and rear moving sliders (410) arranged on the rack (45) and meshingly connected with the rack (45) through a driving motor; The side wall of the left and right moving sliders (44) is slidably connected with the side wall of the toothed rail (43), and the driving motor is arranged in the front and rear moving sliders (410); The motor (46) is controlled by the motion control module; the motion control module comprises a motor driver (48) and a power storage unit; the power storage unit supplies power to the motor driver (48); the motor driver (48) is electrically connected with the processing unit (22); and the motor driver (48) is fixedly arranged below the front and rear moving sliders (410).

2. The adjustable greenhouse crop phenotyping image acquisition and environmental monitoring control device according to claim 1, characterized in that, The image acquisition module (1) comprises a camera (11); the model of the camera (11) is Azure Kinect depth camera; and the camera (11) is electrically connected with the processing unit (22).

3. The adjustable greenhouse crop phenotyping image acquisition and environmental monitoring control device according to claim 1, characterized in that, The environmental parameter acquisition module (3) comprises an illumination intensity sensor (31), an air temperature and humidity sensor (32) and a soil temperature and humidity sensor (33); the illumination intensity sensor (31), the air temperature and humidity sensor (32) and the soil temperature and humidity sensor (33) are arranged on the support columns (41) through mounting rings (47).

4. The adjustable greenhouse crop phenotyping image acquisition and environmental monitoring control device according to claim 1, characterized in that, The support post (41) is a support post with a telescopic rod (49); the upper end of the telescopic rod (49) is connected with the cross beam (42).

5. The adjustable greenhouse crop phenotyping image acquisition and environmental monitoring control device according to claim 4, characterized in that, The communication unit (21) comprises a wireless transmission WIFI module, and the processing unit (22) comprises a Raspberry Pi.

6. The adjustable greenhouse crop phenotyping image acquisition and environmental monitoring control device according to claim 5, characterized in that, The communication unit (21) comprises an extended Ethernet interface.

Images