CN115407771A

CN115407771A - Crop monitoring method, system and device based on machine vision

Info

Publication number: CN115407771A
Application number: CN202210955766.XA
Authority: CN
Inventors: 唐睿智; 王颖琳; 谢泽文; 黄永柱; 李政驰; 吴晓杰; 陈仲坚; 杨世浩; 陈嘉涛; 卢麒霖
Original assignee: Guangzhou University
Current assignee: Guangzhou University
Priority date: 2022-08-10
Filing date: 2022-08-10
Publication date: 2022-11-29

Abstract

The invention relates to the field of agriculture, and discloses a crop monitoring method, a crop monitoring system and a crop monitoring device based on machine vision, wherein the method comprises the following steps: constructing a database of normal crops in different seasons; training a model; performing coarse identification through a main body vision module in the vision modules; the robot is used for accurately identifying the physiological characteristics of crops and is divided into two conditions of identifying the physiological characteristics of the crops by a vision module and acquiring environmental data by a sensor module; the central calculation module compares the information transmitted by the vision module with a database of the central calculation module, judges whether the crop is abnormal or not, calculates the abnormal grade, and obtains the water quantity, the chemical fertilizer quantity and the herbicide quantity of the crop needing irrigation after processing the data. The system includes a patrol module, a sensor module, a processing module, a vision module, a communication module, and a central computing module. The method is based on plant physiology knowledge, different models are constructed according to different environmental conditions of crops, and whether the models are abnormal or not is judged by utilizing a machine vision technology.

Description

Crop monitoring method, system and device based on machine vision

Technical Field

The invention relates to the field of agriculture, in particular to a crop monitoring method, a crop monitoring system and a crop monitoring device based on machine vision.

Background

The existing centralized agricultural planting base generally adopts artificial cultivation and artificial detection of various parameters, the artificial method has low efficiency, but the needs of crops are increased day by day. With the increasing perfection of biological knowledge systems and the vigorous development of automation control technology and machine vision technology, a new method for collecting and monitoring data based on biological knowledge and combining with the machine vision technology is designed, so that the planting planning, management and control of crops can be better performed.

With current patent disclose portable node and fixed node monitoring technology:

the fixed nodes comprise temperature sensors, humidity sensors, soil pH value sensors and soil trace element sensors, the sensors are all connected with the wireless transmission module, the fixed nodes are distributed at different positions in the farmland, the soil temperature and humidity, the pH value, the soil mineral element content and the air temperature and humidity of different places in the farmland are collected, and then the wireless transmission module sends data.

The mobile node comprises a robot provided with a raspberry group, an Arduino development board and a wireless transmission module, the robot receives information collected by fixed nodes distributed everywhere in the moving process, intelligent monitoring software is deployed in the raspberry group, a user establishes connection with the raspberry group in a wireless mode, the intelligent monitoring software is accessed, and data information collected by different fixed nodes is checked in real time.

The defects of the existing agricultural monitoring technology are as follows:

the physiological characteristics of crops are influenced by the environment and have different appearance characteristics in different seasons. The prior art only refers to the judgment by using image information roughly, and does not relate to the identification judgment of the characteristics of crops in different seasons.

In dealing with the weed problem, the prior art uses a training model to judge weeds, but does not mention how many doses of herbicide should be selected after identifying weeds.

The blades are dense in the later stage of crop growth, the problem that the blades of crops shield and the problem of lens light taking are difficult to solve in the prior art, and the accuracy and the effect of image recognition are influenced.

In the growth process of crops, the growth environment of the crops generally does not change too much, and the fixed node sensor has high cost and large energy consumption for acquiring real-time data.

Disclosure of Invention

Technical problem to be solved

The invention provides a crop monitoring method, a crop monitoring system and a crop monitoring device based on machine vision, which aim to solve the problems.

(II) technical scheme

Aiming at the defects of the prior art, the invention provides the following technical scheme:

a crop monitoring method based on machine vision comprises the following steps:

the first step is as follows: constructing a database of normal crops in different seasons;

the second step is that: training the models, and selecting the corresponding trained models according to the corresponding seasons;

the third step: performing coarse identification through a main body vision module in the vision modules;

the fourth step: the robot is used for accurately identifying the physiological characteristics of crops and is divided into two conditions of identifying the physiological characteristics of the crops by a vision module and acquiring environmental data by a sensor module;

the fifth step: the central calculation module compares the information transmitted by the vision module with a database of the central calculation module, judges whether the crop is abnormal or not, calculates the abnormal grade, and obtains the water quantity, the chemical fertilizer quantity and the herbicide quantity of the crop needing irrigation after processing the data.

When the visual module identifies that the physiological characteristics of crops are normal, normal data are recorded, corresponding water and chemical fertilizer are applied, data are fed back, when the visual module identifies that the physiological characteristics of crops are abnormal, abnormal plants are marked, and the abnormal plants are judged, and the processing module executes a corresponding processing method.

The sensor module acquires environmental data and feeds the data back to the central computing module, when the central computing module is normal in computing, corresponding water and chemical fertilizers are applied to feed back the data, when the central computing module is abnormal in computing, the data is compared with an original database for analysis, the abnormality is judged, abnormal plants are marked, and processing is carried out.

A crop monitoring system based on machine vision comprises a patrol module, a sensor module, a processing module, a vision module, an auxiliary module, a communication module and a central computing module, wherein the output end of the patrol module is connected with the input end of the sensor module, the output end of the sensor module is connected with the input end of the communication module, the output end of the patrol module is connected with the input end of the vision module, the output end of the vision module is connected with the input end of the communication module, the output end of the communication module is connected with the input end of the central computing module, the output end of the central computing module is connected with the input end of the processing module, and the output end of the auxiliary module is connected with the input end of the vision module.

The utility model provides a crops monitoring devices based on machine vision, includes solar module, vision module, detection module and drive module, and solar module, vision module and detection module all set up on drive module.

Preferably, drive module includes frame, track, drive wheel and CD-ROM drive motor, the inside fixed mounting of frame has CD-ROM drive motor, the drive wheel sets up in the both sides of frame, all is provided with two tracks on the drive wheel of both sides, the region in the middle of the drive wheel is the thrust wheel, and CD-ROM drive motor links together with the drive wheel, and drive motor's connecting axle is connected through the bradyseism ware between connecting axle and the frame to the drive wheel, installs temperature sensor and carbon dioxide sensor on the frame.

Preferably, detection device includes motor three, lead screw, slider and sensor module, the bottom fixed mounting of frame has motor three, fixed mounting has the lead screw on the output of motor three, the regional threaded connection in below of lead screw and frame is in the same place, the below movable mounting of lead screw has the slider, be provided with sensor module on the slider, sensor module includes soil pH valve sensor and humidity transducer.

Preferably, the vision module comprises a claw, a first mechanical arm, a camera, a second mechanical arm, a chip and a holder, the claw is arranged at the free end of the first mechanical arm, the chip, the holder, the claw and the second mechanical arm are all arranged on the frame, and the free end of the camera is arranged on the second mechanical arm.

Preferably, the solar module includes transmission shaft six, motor two, transmission shaft seven, motor one, transmission shaft one, conveyer belt, transmission shaft two, connecting rod, transmission shaft three, solar panel, helical gear one, helical gear two, transmission shaft four, gear one, transmission shaft five and gear two, the inside fixed mounting of frame has motor one, equal fixed mounting has transmission shaft one on the output shaft of motor both sides, transmission shaft one is connected transmission shaft two through the conveyer belt, is provided with gear one on the frame, transmission shaft two is connected with gear two, and gear two meshes with gear one, and gear one is connected transmission shaft three, the last connecting rod of installing of solar panel installs transmission shaft four and transmission shaft five on the connecting rod, helical gear two is installed to transmission shaft four's one end, helical gear one is installed to the one end of transmission shaft three, helical gear one meshes with helical gear two.

(III) advantageous effects

Compared with the prior art, the crop monitoring method, the crop monitoring system and the crop monitoring device based on machine vision have the following beneficial effects:

1. the invention relates to a crop monitoring method, a crop monitoring system and a crop monitoring device based on machine vision.A different model is constructed according to different environmental conditions of crops based on plant physiology knowledge, the physiological characteristics of the crops under the different environmental conditions and the physiological characteristics of the crops in different growth stages are recorded, and the obtained model is used for constructing a database, and image recognition is carried out in the growth process of the crops by utilizing the machine vision technology to judge whether the crops are abnormal or not. The micro camera with high degree of freedom and sensitivity is used for monitoring, and the problem that the blades are shielded mutually is solved. In the division of the agricultural production area, the robots in different areas mark weeds and the number of the weeds while identifying the weeds. And measuring environmental data regularly by using a sensor carried by the robot, and then determining the water quantity, the chemical fertilizer quantity and the herbicide quantity to be irrigated.

2. The crop monitoring method, the crop monitoring system and the crop monitoring device based on the machine vision are characterized in that a model for judging whether crops are abnormal or not is established, the model can be used for judging and identifying in different growth periods and different seasonal environments of the crops, environmental data obtained by monitoring in regular time intervals are used for calculating the water quantity, chemical fertilizer quantity and herbicide quantity for irrigating marked crops, a multi-machine cooperative operation system is established, the problem of blade shading is solved, the crops are monitored in the regular time intervals according to the physiological cycle of specific crops, and the effects of reducing energy consumption and reducing cost are achieved.

3. According to the crop monitoring method, system and device based on machine vision, an original normal-growth crop model is constructed, whether the crop is abnormal or not is judged according to physiological characteristics of the crop when the crop is diseased, and accuracy is improved.

4. The crop monitoring method, the crop monitoring system and the crop monitoring device based on the machine vision acquire environmental data of specified crops according to the sensors, and perform proper water irrigation and herbicide spraying on the crops.

5. According to the crop monitoring method, the crop monitoring system and the crop monitoring device based on the machine vision, regular monitoring time periods are divided according to the growth characteristics of crops and different environmental conditions, and energy consumption is reduced.

6. The crop monitoring method, the crop monitoring system and the crop monitoring device based on the machine vision utilize the miniature camera with high degree of freedom and sensitivity to monitor, and solve the problem that the blades are shielded mutually.

7. According to the crop monitoring method, system and device based on machine vision, a novel patrol robot is designed, the novel patrol robot is suitable for complex environments, a novel solar rotating device is designed, and the solar energy is converted more efficiently and an innovative sensor monitoring structure carried on the patrol robot is realized.

Drawings

FIG. 1 is a schematic flow chart of a monitoring method according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of a monitoring system according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of another exemplary embodiment of a monitoring system;

FIG. 4 is a schematic structural diagram of an assembled axle side of the patrol module according to an embodiment;

FIG. 5 is a schematic diagram of an assembled patrol module according to an embodiment of the present disclosure;

FIG. 6 is a left side view of a driver module in the patrol module according to an embodiment;

fig. 7 is a schematic diagram of an axis structure of a driving module in the patrol module according to the embodiment.

In the figure: 1. a claw; 2. a first mechanical arm; 3. a camera; 4. a second mechanical arm; 5. a chip; 6. a holder; 7. a temperature sensor; 8. a carbon dioxide sensor; 9. a sixth transmission shaft; 10. a second motor; 11. a transmission shaft seven; 12. a first motor; 13. a first transmission shaft; 14. a conveyor belt; 15. a second transmission shaft; 16. a connecting rod; 17. a third transmission shaft; 18. a solar panel; 19. a first bevel gear; 20. a second bevel gear; 21. a fourth transmission shaft; 22. a first gear; 23. a fifth transmission shaft; 24. a second gear; 25. a third motor; 26. a screw rod; 27. a slider; 28. a soil pH sensor; 29. a humidity sensor; 30. a crawler belt; 31. a drive wheel; 32. a thrust wheel; 33. a shock absorber; 34. the motor is driven.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.

Examples

Referring to fig. 1 to 7, a crop monitoring method based on machine vision according to an embodiment of the present invention includes the following steps:

the fourth step: carrying out accurate identification by a robot;

the fifth step: the central computing module compares the information transmitted by the vision module with a database of the central computing module, judges whether the crops are abnormal or not and computes the abnormal grade, and obtains the water quantity, the fertilizer quantity and the herbicide quantity of the crops needing irrigation after processing the data.

And in the fourth step, the robot is used for accurately identifying the physiological characteristics of the crops by the vision module and acquiring the environmental data by the sensor module.

Referring to fig. 2-3, a crop monitoring system based on machine vision includes a patrol module, a sensor module, a processing module, a vision module, an auxiliary module, a communication module and a central computing module, wherein an output end of the patrol module is connected with an input end of the sensor module, an output end of the sensor module is connected with an input end of the communication module, an output end of the patrol module is connected with an input end of the vision module, an output end of the vision module is connected with an input end of the communication module, an output end of the communication module is connected with an input end of the central computing module, an output end of the central computing module is connected with an input end of the processing module, and an output end of the auxiliary module is connected with an input end of the vision module.

The vision module comprises a main body vision module and a mobile vision module, wherein the vision detection system adopts a CCD camera to convert a detected target into image signals and transmits the image signals to a special image processing system, and the image processing system performs various operations on the signals to extract the characteristics of the target according to the information of pixel distribution, brightness, color and the like and then compares the extracted characteristics with a self database to judge whether the target is abnormal or not; the main part vision module is responsible for carrying out the coarse detection to crops, removes the vision module and carries out the accurate detection to concrete certain crops to it adopts the degree of freedom and the high miniature camera of sensitivity, realizes that the multi-angle acquires the image.

The robot collects soil humiture, pH value, soil mineral element content, air humiture and carbon dioxide content of different places in the farmland at different positions, and then the remote desktop monitors the intelligent monitoring software through the communication module.

And the communication module ensures data and signal transmission between other modules.

The central computing module is composed of a processor chip, and after the module acquires data information transmitted by other modules, the module carries out computing processing and system modeling on the data information, compares the data information with a database of the module according to the information transmitted by the visual module, judges whether the crops are abnormal or not and computes the abnormal grade, and obtains the water quantity, the chemical fertilizer quantity and the herbicide quantity of the crops needing irrigation after processing the data.

And the processing module adopts corresponding measures in the marked crops.

A vision module:

the main part camera: after the camera is used for obtaining the image, the yolov4 target detection algorithm is used for identifying the characteristics of leaves, soil, stems, fruits and the like of crops according to the trained model, and the identified data are transmitted to the central computing module through the communication module.

Mobile robot camera: aiming at the difficulty in identifying images which cannot be identified by a main camera or the difficulty in identifying the overlapped blades, the patrol robot carrying the miniature camera with high freedom degree and sensitivity is used for monitoring, and the problem of difficulty in identifying the overlapped blades is solved.

A central computing module:

training the health and abnormal model characteristics: the crop identified by the vision module is processed, including but not limited to the following features:

blade shape

Whether the size of the blade is normal or not; whether the leaves have a withered, necrotic or rotten condition.

Color change of blade

Whether the color of each part of the leaf (such as the leaf vein and the leaf margin) is normal green (such as whether the leaf has the conditions of yellowing, reddening and the like); the green color of the leaves expresses the physiological status of the leaves to different degrees (for example, the dark green leaves are mature leaves);

spots or lesions

Spots or lesions with different shapes, sizes and colors are formed on the leaves, stems and fruits.

Mildew-like substance

The soil mildew identification agent is composed of hypha of fungi, is usually attached to soil, and identifies whether the color, shape and structure of a mildew layer are abnormal to those of soil with normal leaves.

Weed (Haw)

Applying different herbicides to different weeds

Planning regular monitoring time intervals according to the physiological cycle of crops:

the system communication network judges the current season according to the current geographical position and time and the model difference under the normal state corresponding to different seasons, calls the corresponding model according to the current season system to improve the accuracy, and applies different amounts of water and fertilizer under the conditions of different seasons for different crops

Calculating the herbicide dosage required to be sprayed according to the weed quantity:

and constructing a corresponding identification model according to common weeds, comparing the identification model with a database when a certain weed is identified, and outputting a corresponding weed dose to achieve the optimal effect.

A sensor module:

including but not limited to temperature sensors, humidity sensors, soil ph sensors, soil trace element sensors, etc., all sensors that facilitate the determination accuracy of the present system are within the contemplation of the present inventors.

A communication module:

by the common communication method

An execution module:

the manual process is notified by a buzzer or the like producing a sound.

An auxiliary module:

the light source is used for supplementing light and common image processing algorithms such as night vision algorithm and defogging algorithm are used for assisting visual recognition, the photovoltaic panel is carried, solar charging is achieved, and the working time is prolonged.

Referring to fig. 4-7, a crop monitoring device based on machine vision includes a solar module, a vision module, a detection module and a driving module, wherein the solar module, the vision module and the detection module are all disposed on the driving module.

Further, the driving module comprises a frame, two crawler belts 30, a driving wheel 31 and a driving motor 34, the driving motor 34 is fixedly mounted inside the frame, the driving wheel 31 is arranged on two sides of the frame, the two crawler belts 30 are arranged on the driving wheels 31 on two sides, a supporting wheel 32 is arranged in the middle of the driving wheel 31, the driving motor 34 is connected with the driving wheel 31, a connecting shaft of the driving wheel 31 connected with the driving motor 34 is connected with the frame through a shock absorber 33, and the frame is provided with a temperature sensor 7 and a carbon dioxide sensor 8.

Further, the detection device comprises a third motor 25, a lead screw 26, a sliding block 27 and a sensor module, the third motor 25 is fixedly mounted at the bottom of the frame, the lead screw 26 is fixedly mounted at the output end of the third motor 25, the lead screw 26 is in threaded connection with the lower area of the frame, the sliding block 27 is movably mounted below the lead screw 26, the sensor module is arranged on the sliding block 27 and comprises a soil pH value sensor 28 and a humidity sensor 29.

Further, the vision module comprises a claw 1, a first mechanical arm 2, a camera 3, a second mechanical arm 4, a chip 5 and a cloud platform 6, wherein the claw 1 is arranged at the free end of the first mechanical arm 2, the chip 5, the cloud platform 6, the claw 1 and the second mechanical arm 4 are all arranged on the frame, and the free end of the camera 3 is arranged on the second mechanical arm 4.

Further, the solar module comprises a transmission shaft six 9, a motor two 10, a transmission shaft seven 11, a motor one 12, a transmission shaft one 13, a conveyor belt 14, a transmission shaft two 15, a connecting rod 16, a transmission shaft three 17, a solar panel 18, a helical gear one 19, a helical gear two 20, a transmission shaft four 21, a gear one 22, a transmission shaft five 23 and a gear two 24, wherein the motor one 12 is fixedly installed inside the frame, the transmission shaft one 13 is fixedly installed on output shafts on two sides of the motor one 12, the transmission shaft one 13 is connected with the transmission shaft two 15 through the conveyor belt 14, the frame is provided with the gear one 22, the transmission shaft two 15 is connected with the gear two 24, the gear two 24 is meshed with the gear one 22, the gear one 22 is connected with the transmission shaft three 17, the connecting rod 16 is installed on the solar panel 18, the transmission shaft four 21 and the transmission shaft five 23 are installed on the connecting rod 16, the transmission shaft four 21 is provided with the helical gear two 20, one end of the transmission shaft three 17 is provided with the helical gear one 19, and the helical gear one 19 is meshed with the helical gear two 20.

The solar module principle is as follows: the chip 5 controls the solar panel 18 to open the panel whenever the device temporarily needs charging or the device is idle. Firstly, after the first motor 12 receives the instruction of the chip 5, two ends of the first motor 12 rotate at the same rotation speed and in opposite directions for a certain angle: the first motor 12 transmits power to the first transmission shaft 13, the second transmission shaft 15 transmits power from the first motor 14, the first gear 22 and the second gear 24 reversely rotate and then transmit the power to the third transmission shaft 17, the power of the second transmission shaft 15 and the third transmission shaft 17 is converted from rotation on a side vertical plane to rotation on a normal vertical plane through the first bevel gear 19 and the second bevel gear 20, and the fourth transmission shaft 21 and the fifth transmission shaft 23 are driven to rotate. Since the power is reversed in the front, the rotation direction of the transmission shaft four 31 is the same as that of the transmission shaft five 23, and the connecting rod 16 is driven to rotate by the double shafts. Meanwhile, the second motor 10 rotates for a certain angle to drive the sixth transmission shaft 9 and the seventh transmission shaft 11 to realize the rotation of the solar panel. The two processes work in cooperation to control the elevation of the solar panel 18.

The visual module works on the following principle: the claw 1 is arranged on the first mechanical arm 2, the camera 3 is arranged on the second mechanical arm 4, and the two mechanical arms are respectively arranged on the cloud deck 6. The camera 3 can move in all directions through the movement of the mechanical arm and the three steering engines on the tripod head 6. When the camera 3 is blocked by other plants, the chip 6 controls the second mechanical arm 4 to move, accurately moves to the blocked position and pulls out the blocked plants.

A detection module: the temperature sensor 7 and the carbon dioxide sensor 8 are fixed on the system frame because they do not need to contact with the soil, and the soil pH sensor 28 and the humidity sensor 29 are designed as a lifting device because they contact with the soil. When the system moves to a designated position, the third motor 25 starts to work to drive the screw rod 26 to rotate. Due to the limitation of the frame, the screw rod 26 drives the slide block 27 to do descending translation, after the slide block 27 descends to a set value, the soil pH value sensor 28 and the humidity sensor 29 connected to the two ends of the slide block 27 are contacted with soil, the detection is started, and after the detection is finished, the processes are repeated, and the reset of the slide block 27 is finished.

The driving module principle is as follows: considering that the working road surface environment of the system is soil environment such as farmland, and the like, the use of common wheels can cause unstable running and influence the working state of the system due to potholes, so the crawler 30 is adopted for driving. The driving wheel 31 is designed in the middle of the crawler 30, the driving wheels 31 on two sides are used as the thrust wheels 32, and the independent suspension structure is designed on the bearing shaft, so that the supporting and shock-absorbing effects are achieved.

According to the method, different models are built according to different environmental conditions of crops based on plant physiology knowledge, physiological characteristics of the crops under different environmental conditions and physiological characteristics of the crops in different growth stages are recorded, and whether the crops are abnormal or not is judged by utilizing image recognition in the growth process of the crops through a machine vision technology by using a database built by the obtained models. The micro camera with high degree of freedom and sensitivity is used for monitoring, and the problem that the blades are shielded mutually is solved. In the division of the agricultural production area, the robots in different areas mark weeds and the number of the weeds while identifying the weeds. And measuring environmental data regularly by using a sensor carried by the robot, and then determining the water quantity, the chemical fertilizer quantity and the herbicide quantity to be irrigated.

Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims

1. A crop monitoring method based on machine vision is characterized by comprising the following steps:

the second step: training the models, and selecting the corresponding trained models according to the corresponding seasons;

2. The machine-vision based crop monitoring method of claim 1, characterized in that: when the visual module identifies that the physiological characteristics of the crops are normal, normal data are recorded, corresponding water and chemical fertilizers are applied, data are fed back, when the visual module identifies that the physiological characteristics of the crops are abnormal, abnormal plants are marked, and the abnormal plants are judged, and the processing module executes a corresponding processing method.

3. The machine-vision based crop monitoring method of claim 1, wherein: the sensor module acquires environmental data and feeds the data back to the central computing module, when the central computing module is normal in computing, corresponding water and chemical fertilizers are applied to feed back the data, when the central computing module is abnormal in computing, the data and an original database are compared and analyzed, and the abnormal plants are marked and processed.

4. The crop monitoring system based on machine vision is characterized by comprising a patrol module, a sensor module, a processing module, a vision module, an auxiliary module, a communication module and a central computing module, wherein the output end of the patrol module is connected with the input end of the sensor module, the output end of the sensor module is connected with the input end of the communication module, the output end of the patrol module is connected with the input end of the vision module, the output end of the vision module is connected with the input end of the communication module, the output end of the communication module is connected with the input end of the central computing module, the output end of the central computing module is connected with the input end of the processing module, and the output end of the auxiliary module is connected with the input end of the vision module.

5. The utility model provides a crops monitoring devices based on machine vision, its characterized in that includes solar module, vision module, detection module and drive module, and solar module, vision module and detection module all set up on drive module.

6. The machine-vision based crop monitoring apparatus of claim 5, wherein: drive module includes frame, track (30), drive wheel (31) and CD-ROM drive motor (34), the inside fixed mounting of frame has CD-ROM drive motor (34), drive wheel (31) set up in the both sides of frame, all are provided with two track (30) on drive wheel (31) of both sides, the region in the middle of drive wheel (31) is thrust wheel (32), and CD-ROM drive motor (34) link together with drive wheel (31), is connected through bumper (33) between the connecting axle that drive motor (34) was connected in drive wheel (31) and the frame, installs temperature sensor (7) and carbon dioxide sensor (8) on the frame.

7. The machine-vision based crop monitoring apparatus of claim 6, wherein: detection device includes motor three (25), lead screw (26), slider (27) and sensor module, the bottom fixed mounting of frame has motor three (25), fixed mounting has lead screw (26) on the output of motor three (25), lead screw (26) are in the same place with the regional threaded connection in below of frame, the below movable mounting of lead screw (26) has slider (27), be provided with sensor module on slider (27), sensor module includes soil pH valve sensor (28) and humidity transducer (29).

8. The machine-vision based crop monitoring apparatus of claim 6, wherein: the vision module comprises a claw (1), a first mechanical arm (2), a camera (3), a second mechanical arm (4), a chip (5) and a holder (6), wherein the claw (1) is arranged at the free end of the first mechanical arm (2), the chip (5), the holder (6), the claw (1) and the second mechanical arm (4) are all arranged on a frame, and the free end of the camera (3) is arranged on the second mechanical arm (4).

9. The machine-vision based crop monitoring apparatus of claim 6, wherein: the solar module comprises a transmission shaft six (9), a motor two (10), a transmission shaft seven (11), a motor one (12), a transmission shaft one (13), a conveyor belt (14), a transmission shaft two (15), a connecting rod (16), a transmission shaft three (17), a solar panel (18), a bevel gear one (19), a bevel gear two (20), a transmission shaft four (21), a gear one (22), a transmission shaft five (23) and a gear two (24), wherein the motor one (12) is fixedly installed inside the frame, the transmission shafts one (13) are fixedly installed on output shafts on two sides of the motor one (12), the transmission shaft one (13) is connected with the transmission shaft two (15) through the conveyor belt (14), the frame is provided with the gear one (22), the transmission shaft two (15) is connected with the gear two (24), the gear two (24) is meshed with the gear one (22), the gear one (22) is connected with the transmission shaft three (17), the solar panel (18) is provided with the connecting rod (16), the connection rod (16) is provided with the transmission shaft four (21) and the transmission shaft five (23), one end of the transmission shaft four (21) is provided with the second (20), and one bevel gear (17), the first bevel gear (19) is meshed with the second bevel gear (20).