CN116114683B - Flame weeding machine capable of detecting weed density and crops - Google Patents
Flame weeding machine capable of detecting weed density and crops Download PDFInfo
- Publication number
- CN116114683B CN116114683B CN202211663320.6A CN202211663320A CN116114683B CN 116114683 B CN116114683 B CN 116114683B CN 202211663320 A CN202211663320 A CN 202211663320A CN 116114683 B CN116114683 B CN 116114683B
- Authority
- CN
- China
- Prior art keywords
- flame
- crop
- fire
- weed
- crops
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 241000196324 Embryophyta Species 0.000 title claims abstract description 115
- 238000009333 weeding Methods 0.000 title abstract description 45
- 230000007246 mechanism Effects 0.000 claims abstract description 67
- 238000005507 spraying Methods 0.000 claims abstract description 18
- 230000000694 effects Effects 0.000 claims abstract description 12
- 238000001514 detection method Methods 0.000 claims description 24
- 238000010191 image analysis Methods 0.000 claims description 15
- 238000012545 processing Methods 0.000 claims description 15
- 238000010285 flame spraying Methods 0.000 claims description 14
- 230000002363 herbicidal effect Effects 0.000 claims description 8
- 239000004009 herbicide Substances 0.000 claims description 8
- 238000012360 testing method Methods 0.000 claims description 8
- 238000013527 convolutional neural network Methods 0.000 claims description 7
- 238000012549 training Methods 0.000 claims description 6
- 238000005286 illumination Methods 0.000 claims description 5
- 238000000034 method Methods 0.000 claims description 5
- 230000009471 action Effects 0.000 claims description 4
- 238000002372 labelling Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 238000005452 bending Methods 0.000 claims description 3
- 238000005516 engineering process Methods 0.000 claims description 3
- 238000013508 migration Methods 0.000 claims description 3
- 230000005012 migration Effects 0.000 claims description 3
- 230000008859 change Effects 0.000 abstract description 4
- 239000002737 fuel gas Substances 0.000 abstract description 3
- 238000013461 design Methods 0.000 abstract description 2
- 239000007789 gas Substances 0.000 description 32
- 230000001276 controlling effect Effects 0.000 description 8
- 238000009434 installation Methods 0.000 description 5
- 239000000126 substance Substances 0.000 description 4
- 239000002689 soil Substances 0.000 description 3
- RNFJDJUURJAICM-UHFFFAOYSA-N 2,2,4,4,6,6-hexaphenoxy-1,3,5-triaza-2$l^{5},4$l^{5},6$l^{5}-triphosphacyclohexa-1,3,5-triene Chemical compound N=1P(OC=2C=CC=CC=2)(OC=2C=CC=CC=2)=NP(OC=2C=CC=CC=2)(OC=2C=CC=CC=2)=NP=1(OC=1C=CC=CC=1)OC1=CC=CC=C1 RNFJDJUURJAICM-UHFFFAOYSA-N 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 239000003063 flame retardant Substances 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 206010059866 Drug resistance Diseases 0.000 description 1
- 240000008042 Zea mays Species 0.000 description 1
- 235000005824 Zea mays ssp. parviglumis Nutrition 0.000 description 1
- 235000002017 Zea mays subsp mays Nutrition 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002146 bilateral effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 210000002421 cell wall Anatomy 0.000 description 1
- 239000011248 coating agent Substances 0.000 description 1
- 238000000576 coating method Methods 0.000 description 1
- 235000005822 corn Nutrition 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000003028 elevating effect Effects 0.000 description 1
- 238000003912 environmental pollution Methods 0.000 description 1
- 235000013305 food Nutrition 0.000 description 1
- 239000000446 fuel Substances 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000000575 pesticide Substances 0.000 description 1
- 239000000447 pesticide residue Substances 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 230000001850 reproductive effect Effects 0.000 description 1
- 238000004162 soil erosion Methods 0.000 description 1
- 230000008685 targeting Effects 0.000 description 1
Classifications
-
- A—HUMAN NECESSITIES
- A01—AGRICULTURE; FORESTRY; ANIMAL HUSBANDRY; HUNTING; TRAPPING; FISHING
- A01M—CATCHING, TRAPPING OR SCARING OF ANIMALS; APPARATUS FOR THE DESTRUCTION OF NOXIOUS ANIMALS OR NOXIOUS PLANTS
- A01M21/00—Apparatus for the destruction of unwanted vegetation, e.g. weeds
- A01M21/04—Apparatus for destruction by steam, chemicals, burning, or electricity
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
- G06T1/0014—Image feed-back for automatic industrial control, e.g. robot with camera
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/188—Vegetation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Multimedia (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Zoology (AREA)
- Robotics (AREA)
- Environmental Sciences (AREA)
- Wood Science & Technology (AREA)
- Pest Control & Pesticides (AREA)
- Insects & Arthropods (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Catching Or Destruction (AREA)
Abstract
The invention discloses a flame weeding machine capable of detecting weed density and crops, which comprises a moving platform, a weed density and crop identification device, a leaf gathering mechanism, a fire baffle mechanism and a fire spraying device, wherein the moving platform is connected with the weed density and crop identification device; the weed density and the crop identification device are used for identifying weed density information and crop size information on a walking path, and a flaming system is used for adjusting flame intensity according to the identified weed density so as to reasonably utilize fuel gas; the leaf gathering mechanism and the fire baffle mechanism adjust the fire-blocking protection space for the crops according to the size information of the identified crops, and can improve the area for weeding the flames of the crops while playing a good fire-blocking protection role. In addition, the invention also designs a lifting mechanism for controlling the overall ground clearance of the weed density and crop identification device, the leaf gathering mechanism, the fire baffle mechanism and the fire spraying device so as to adapt to the change of terrains and maintain the distance from the flame nozzle to the ground within a reasonable range, thereby ensuring the weeding effect.
Description
Technical Field
The invention belongs to the technical fields of agricultural engineering and agricultural robots, and particularly relates to a flame weeding machine capable of detecting weed density and crops.
Background
Farmland weeds force agricultural producers to invest a great deal of manpower and financial resources on farmland weeding operation every year by virtue of the stubborn vitality and the rapid reproductive capacity of the farmland weeds. In order to save the cost, agricultural producers are urgent to have an efficient and inexpensive weeding method. Under the push of demands, weeding modes such as mechanical weeding, chemical weeding, microwave weeding, electric weeding, laser weeding, flame weeding and the like begin to appear and are gradually optimized. The mechanical weeding has higher requirements on the row spacing and plant spacing of crops, and the mechanical structure has weaker adaptability and is easy to damage the crops, and the soil structure can be influenced to cause soil erosion; the chemical weeding operation is simple, the pertinence is strong, the weeding effect is good, but the long-term use can lead to the accumulation of soil pesticide residues, and the food safety problem is caused. And the drug resistance of weeds is improved, and the use cost of pesticides is increased. The realization of the functions of microwave weeding, electric weeding and laser weeding depends on expensive high-power equipment, and has good weeding effect, but the purchase cost is difficult to balance, so that the large-scale popularization is not facilitated.
Flame weeding rapidly expands liquid inside weeds through high-temperature flame, so that cell walls are broken, and weed leaves die after a period of time. The weeding mode does not affect the soil structure or generate harmful substances, and the weeding cost is considered while the defects of mechanical weeding and chemical weeding are well overcome. Flame weeding is therefore a suitable choice. However, the traditional flame weeding mode has higher requirements on fire resistance of crops, weeding operation can be performed only in a specific crop growth period, and the flame weeding mode causes great waste of flame fuel. With the development of the mechanical field, flame weeding modes with the targeting of machine vision begin to emerge and develop rapidly.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a flame weeding machine capable of detecting weed density and crops.
The invention is realized by the following technical scheme:
A flame weeding machine capable of detecting weed density and crops comprises a moving platform, a weed density and crop identification device, a leaf gathering mechanism, a fire baffle mechanism, a fire spraying device and a lifting mechanism;
The weed density and crop identification device is positioned at the bottom of the mobile platform and comprises a shading black box, an infrared camera, an LED lamp and an image analysis processing system, wherein the infrared camera, the LED lamp and the image analysis processing system are arranged in the shading black box; two layers of shading curtains are respectively arranged on the front side and the rear side of the shading black box at a certain interval;
the leaf gathering mechanism is arranged at two sides of the rear side outlet of the shading black box and comprises leaf gathering plates and first electric telescopic rods used for controlling the opening angles of the leaf gathering plates, and the opening angles of the leaf gathering plates are controlled through the first electric telescopic rods, so that the space clamped between the two leaf gathering plates is matched with the size of crops;
The fire baffle mechanism is arranged at the rear of the leaf gathering mechanism and comprises two fire baffles and a second electric telescopic rod for controlling the opening angle of the fire baffles, the fire baffles are hinged on the leaf gathering plate through hinges, the second electric telescopic rod is arranged between the fire baffles and the leaf gathering plate, and the bottom end of the fire baffles is provided with an inward bending plate;
The fire spraying device is arranged at two outer sides of the fire baffle mechanism and is connected with a fire spraying control system, and the fire spraying control system adjusts the gas flow according to the density of weeds and the density information of the weeds identified by the crop identification device;
The lifting mechanism is used for controlling the weed density and the overall ground clearance of the crop identification device, the leaf gathering mechanism, the fire baffle mechanism and the fire spraying device, and the distance from the flame nozzle to the ground is maintained within a reasonable range.
In the technical scheme, weeds and crops can lift the shade to enter/exit the shading black box in the travelling process of the mobile platform, the shading effect inside the shading black box is guaranteed under the effect of the shade, and the images of the weeds and the crops passing through the inside of the shading black box are collected under the auxiliary illumination effect of the infrared camera and the LED lamp in the shading black box; the collected images are input into an image analysis processing system, and the density of weeds and the size information of crops are identified by the image analysis processing system.
In the technical scheme, the leaf gathering mechanism controls the space clamped between the two leaf gathering plates to be matched with the size of the crop according to the weed density and the size information of the crop identified by the crop identification device.
In the technical scheme, the baffle is arranged on the inner side of the leaf gathering plate and used for gathering the crop leaves in a lifting mode, and the angle of the baffle is adjusted through the electric telescopic rod.
In the technical scheme, the flame spraying device comprises a connecting rod and a flame spraying device, wherein the flame spraying device is arranged on the outer side wall of the flame baffle plate through the connecting rod; the flame thrower comprises a flame shielding cover, a plurality of flame nozzles are arranged on the inner side of the flame shielding cover, and the flame nozzles are connected to a flame spraying control system through pipelines.
In the above technical scheme, the flaming control system comprises a gas tank, a gas booster pump and a gas flow controller, wherein the gas inlet of the gas flow controller is connected with the gas outlet of the gas booster pump through a high-pressure hose, and the gas outlet of the gas flow controller is communicated with a flame nozzle of the flaming device through a pipeline.
In the above technical scheme, set up fore-and-aft mounting groove on the flame retardant coating, the top of connecting rod passes through bolt fixed mounting on the mounting groove, through the mounted position of the top of adjusting the connecting rod on the mounting groove, can adjust the initial height of flame thrower apart from ground.
In the above technical scheme, elevating system includes four vertical electric lifter, is the rectangle and distributes, installs moving platform bottom, four vertical electric lifter bottom are connected weed density and crop recognition device's shading black box.
In the technical scheme, the deep-sinking wheel supporting mechanisms are arranged on two sides of the shading black box and comprise elastic telescopic rods and deep-sinking wheel supports, the top ends of the elastic telescopic rods are fixedly arranged on the shading black box, the bottom ends of the elastic telescopic rods are connected with the center positions of the tops of the deep-sinking wheel supports, the deep-sinking wheel supports are provided with two supporting legs, and the deep-sinking wheels are arranged at the bottoms of the supporting legs; the elastic telescopic rod provides downward elastic restoring force for the deep-sinking wheel support, so that the deep-sinking wheel can be always tightly attached to the ground.
In the technical scheme, the image analysis processing system adopts a model based on a YOLO V4-tiny convolutional neural network to identify weeds and crops; the steps of establishing the convolutional neural network model for identifying weeds and crops and calculating weed density and crop area are as follows:
Step 1: collecting infrared images of field weeds and crops through the mobile platform, labeling the images, dividing the labels into crops and weeds, and obtaining a data set and corresponding label files after all the images are labeled;
step 2: dividing the data set into a training data set and a testing data set, utilizing the data enhancement expansion data set, combining a migration learning technology to train the YOLO V4-tiny model and optimizing model weights;
Step 3: predicting field weeds and crop images by using the model trained in the step 2, comparing the field weeds and crop images with real weeds and crop information in the test data set, quantifying prediction errors, and evaluating model accuracy;
step 4: identifying a target image acquired in the shading black box by adopting the optimal model evaluated in the step 3, and identifying a weed detection frame and a crop detection frame contained in the target image;
Calculating the area of each weed detection frame contained in the target image according to the recognized frame length and width of each weed detection frame and the center point coordinates, summing the areas of all the weed detection frames contained in the target image, removing the overlapping areas of the detection frames, and obtaining the coverage area of weeds in the target image, wherein the coverage area of weeds is divided by the area of the target image, so that the weed density of the target area can be calculated;
And calculating the area of the crops contained in the target image according to the recognized frame length and width of the crop detection frame and the center point coordinates.
The invention has the advantages and beneficial effects that:
According to the invention, the weed density and the crop identification device can efficiently and accurately identify the weed density information and the crop size information on the walking path in real time, and the flame intensity is regulated by the flaming system according to the identified weed density, so that the fuel gas is reasonably utilized, the fuel gas waste under the weed-free and low-density weed state is avoided, and the environmental pollution is reduced; in addition, the invention also provides the leaf gathering mechanism and the fire baffle mechanism, which can adjust the fire-blocking protection space for the crops according to the size information of the identified crops, and can improve the area for weeding the flames of the crops while playing a good fire-blocking protection role. In addition, the invention also designs a lifting mechanism for controlling the overall ground clearance of the weed density and crop identification device, the leaf gathering mechanism, the fire baffle mechanism and the fire spraying device so as to adapt to the change of terrains and maintain the distance from the flame nozzle to the ground within a reasonable range, thereby ensuring the weeding effect.
Drawings
Fig. 1 is a schematic perspective view of a flame herbicide according to the present invention.
FIG. 2 is a schematic diagram showing the connection structure of the weed density and crop identification apparatus, the leaf gathering mechanism, the flame retardant plate mechanism and the flame spraying apparatus according to the present invention.
Fig. 3 is a front cross-sectional view of fig. 2.
Fig. 4 is a schematic view of the structure of the flame thrower of the present invention.
Fig. 5 is a schematic structural view of the lifting mechanism in the present invention.
Other relevant drawings may be made by those of ordinary skill in the art from the above figures without undue burden.
Detailed Description
In order to make the person skilled in the art better understand the solution of the present invention, the following describes the solution of the present invention with reference to specific embodiments.
Example 1
Referring to fig. 1-5, a flame weeder capable of detecting weed density and crops comprises a moving platform 1, a weed density and crop identification device 2, a leaf gathering mechanism 3, a fire baffle mechanism 4, a fire spraying device 5 and a lifting mechanism 6.
The mobile platform 1 comprises a frame 1.1, a power supply 1.2, universal wheels 1.3 and driving wheels 1.4, wherein the power supply 1.2 is fixedly arranged on the frame 1.1 and provides a required power supply for the whole flame weeding machine; the two driving wheels 1.4 are arranged below the front part of the frame 1.1, and a hub motor is arranged in the frame and is in an integrated structure with the driving wheels; two universal wheels 1.3 are arranged and are arranged below the rear part of the frame 1.1.
Further, the mobile platform 1 is of a bilateral symmetry structure, the left part and the right part are fixed with the connecting rod through the connecting plate, and the width of the mobile platform can be adjusted through manual adjustment so as to adapt to crops with different row spaces.
The weed density and crop identification device 2 is located at the bottom of the mobile platform 1, and specifically, the weed density and crop identification device 2 comprises a shading black box 2.1, an infrared camera 2.2, an LED lamp 2.3 and an image analysis processing system, wherein the infrared camera 2.2, the LED lamp 2.3 and the image analysis processing system are arranged inside the shading black box. Wherein, shading black case 2.1 includes rectangular frame, top cap and controls two curb plates, and the top cap is installed at rectangular frame top, and control two curb plates and install respectively in rectangular frame's left and right sides (rectangular frame's left and right direction is the transverse direction along moving platform), respectively install two-layer window shade 2.4 at rectangular frame's front and back both sides according to certain interval (rectangular frame's front and back direction sets up the front and back direction along moving platform's march), establish two-layer window shade's purpose is: in the travelling process of the mobile platform, weeds and crops can lift the window shade to enter/exit the shading black box, the shading effect inside the shading black box is guaranteed under the action of the window shade, and then the weeds and crops passing through the inside of the shading black box are subjected to high-quality and stable image acquisition under the auxiliary illumination action of the infrared camera and the LED lamp inside the shading black box; the collected images are input into an image analysis processing system, and the image analysis processing system analyzes the density of weeds and the size information of crops.
In addition, it is to be noted that, respectively according to certain interval installation two-layer window shade in both sides around the shading black box, just can guarantee that the crop passes through behind the first layer window shade and just contacts with the second floor window shade, when avoiding the crop to pass through the window shade with the window shade lift, lead to external light source to penetrate into the shading black box inside, cause illumination environment to change drastically, influence the image acquisition effect, namely: the shading curtain is matched with the shading black box to isolate unstable illumination environment. Further, the total number of the LED lamps 2.3 is four, and the LED lamps are respectively and fixedly arranged on four walls inside the shading black box 2.1 to provide a stable light source in a shading environment.
The leaf gathering mechanism 3 is arranged at two sides of the rear side outlet of the shading black box 2.1 and is used for gathering leaves of crops according to the density of weeds and the size information of the crops identified by the crop identification device 2, and protecting the crops in subsequent flaming weeding work. Specifically, the leaf gathering mechanism 3 comprises leaf gathering plates 3.1 and first electric telescopic rods 3.2 for controlling the opening angles of the leaf gathering plates 3.1, the number of the leaf gathering plates 3.1 is two, the leaf gathering plates 3.1 are symmetrically arranged on two sides of a rear side outlet of the shading black box 2.1, the leaf gathering plates 3.1 are hinged with the shading black box 2.1 through hinges, the first electric telescopic rods 3.2 are installed between the leaf gathering plates 3.1 and the shading black box 2.1, and then the opening angles of the leaf gathering plates 3.1 are controlled through the first electric telescopic rods 3.2, so that the space clamped between the two leaf gathering plates is matched with the size of crops, and a good leaf gathering effect is achieved. Further, a trapezoid board m (see fig. 1) forming an angle of 35 degrees with the ground is installed on the inner side of the leaf gathering board, so as to lift up and gather the crop leaves, and the angle of the trapezoid board can be adjusted through an electric telescopic rod (not shown in the drawing) so as to adapt to crops in different growth periods.
The fire baffle mechanism 4 is arranged behind the leaf gathering mechanism 3 and has the function of well protecting crops from fire, in particular preventing the rear side of the crops from being burnt by flame. Specifically, the fire baffle mechanism 4 comprises two fire baffles 4.1 and a second electric telescopic rod 4.2 for controlling the opening angle of the fire baffles, the fire baffles 4.1 are hinged on the gathering leaves 3.1 through hinges, and the second electric telescopic rod 4.2 is arranged between the fire baffles 4.1 and the gathering leaves 3.1, so that the included angle between the fire baffles 4.1 and the gathering leaves 3.1 can be adjusted; the bottom end of the fire baffle 4.1 is provided with an inward bending plate n, so that gathered blades can be prevented from sliding off, the weeding area of flames on crops is increased, and the flames are prevented from entering the space clamped by the gathering mechanism 3 from the rear as much as possible.
The flame spraying device 5 is arranged at two outer sides of the fire baffle mechanism 4 and is used for spraying flame to play a role in weeding the flame. Specifically, the flame spraying device 5 comprises a connecting rod 5.1 and a flame spraying device 5.2, wherein the flame spraying device 5.2 is arranged on the outer side wall of the flame baffle 4.1 through the connecting rod 5.1. Referring to fig. 4, the flame thrower 5.2 includes a flame guard 5.21, a plurality of flame nozzles 5.22 are disposed inside the flame guard 5.21 (4 flame nozzles 5.22 on each flame guard 5.21, three flame directions being perpendicular to the ground and one flame direction being at an angle to the ground), and the flame nozzles 5.22 are connected to a flame control system through a pipe. The flaming control system comprises a gas tank 7, a gas booster pump 8 and a gas flow controller 9, wherein the gas tank 7 is fixedly arranged on the mobile platform 1 and is used for storing gas; the gas booster pump 8 is arranged on the mobile platform 1 and positioned at one side of the gas tank 7, and is used for providing stable gas release pressure for the flaming device 5 and ensuring stable gas flow; the gas flow controller 9 is fixedly arranged on the outer wall of the blade-gathering plate 3.1 and used for accurately measuring and controlling the gas flow, the problem of unstable flame intensity caused by pressure drop of the gas tank 7 along with gas consumption is avoided, the gas inlet of the gas flow controller 9 is connected with the gas outlet of the gas booster pump 8 through a high-pressure hose, and the gas outlet of the gas flow controller 9 is communicated with the flame nozzle 5.22 of the flame spraying device 5 through a pipeline. In operation, the gas flow controller 9 adjusts the gas flow rate based on the weed density and the density information of the weeds identified by the crop identification apparatus 2, and the gas flow rate increases as the weed density increases.
Further, a longitudinal installation groove a (see fig. 3) is formed in the fire baffle 4.1, the top of the connecting rod 5.1 is fixedly installed on the installation groove through a bolt, and the initial height of the flame thrower 5.2 from the ground can be adjusted by adjusting the installation position of the top of the connecting rod 5.1 on the installation groove.
The lifting mechanism 6 is used for controlling the overall ground clearance of the weed density and crop identification device 2, the leaf gathering mechanism 3, the fire baffle mechanism 4 and the flaming device 5 so as to adapt to the change of terrains and maintain the distance from the flame nozzle to the ground within a reasonable range, thereby ensuring the weeding effect. Specifically, referring to fig. 1 and fig. 5, the lifting mechanism 6 includes four longitudinal electric lifting rods, which are rectangular in distribution and are installed at the bottom of the moving platform, the bottom ends of the four longitudinal electric lifting rods are connected with the shading black box 2.1 of the weed density and crop identification device 2, so that the weed density and crop identification device 2 can be driven to move up and down, and the leaf gathering mechanism 3, the fire baffle mechanism 4 and the fire spraying device 5 are integrated with the shading black box 2.1 of the weed density and crop identification device 2, so that the lifting mechanism 6 can synchronously drive the weed density and crop identification device 2, the leaf gathering mechanism 3, the fire baffle mechanism 4 and the fire spraying device 5 to move up and down, and the whole regulator is in a gap between the ground. During adjustment, the lifting mechanism 6 is adjusted according to the detected height information of the flaming device 5 from the ground, and the information can be obtained through measurement by a distance measuring sensor or analysis by using the weed density and the image acquired by the crop identification device 2.
Further, deep-sinking wheel supporting mechanisms 10 are arranged on two sides of the shading black box 2.1, the deep-sinking wheel supporting mechanisms 10 comprise elastic telescopic rods 10.1 and deep-sinking wheel supports 10.2, the top ends of the elastic telescopic rods 10.1 are fixedly arranged on the shading black box 2.1, the bottom ends of the elastic telescopic rods 10.1 are connected with the center position of the top of the deep-sinking wheel supports 10.2, the deep-sinking wheel supports 10.2 are provided with two supporting legs, and deep-sinking wheels 10.3 are arranged at the bottoms of the supporting legs; the elastic telescopic rod 10.1 provides downward elastic restoring force for the deep-sinking wheel support 10.2, so that the deep-sinking wheel 10.3 can be always tightly attached to the ground, the working pressure of the lifting mechanism 6 is relieved, and meanwhile, the whole stability of the shading black box 2.1, the leaf gathering mechanism 3 connected with the shading black box 2.1, the fire baffle mechanism 4 and the fire spraying device 5 is guaranteed, and severe shaking in the lifting process is avoided. Specifically, the elastic telescopic rod 10.1 comprises an outer cylinder, an inner rod and a spring, wherein the inner rod is inserted in the outer cylinder, the spring is arranged between the outer cylinder and the inner rod, downward elastic force is provided for the inner rod through the spring, the lower end of the inner rod is connected with the deep-sinking wheel support 10.2, and further downward elastic restoring force is provided for the deep-sinking wheel support 10.2.
Example two
Furthermore, the flame weeding machine adopts an autonomous navigation system, and the autonomous navigation system comprises an RGB camera 11 and a GNSS receiver 12, wherein the RGB camera 11 is arranged at the middle position in front of the mobile platform 1, and is used for collecting field images in the advancing direction of the flame weeding machine, identifying crop ridge lines by an autonomous navigation algorithm in an upper computer and further extracting navigation lines. The GNSS receiver is symmetrically and fixedly installed at the two sides of the front end of the mobile platform 1 by taking the RGB camera as a center, acquires the position information of the weeding machine in real time through Beidou navigation positioning, and realizes the autonomous mobile path planning of the weeding machine by combining the navigation line. And the obtained information is sent to an automatic driving control module to control the rotation speed of the driving wheel 1.4 so as to realize start and stop, movement and turning.
Example III
Further, the image analysis processing system adopts a model based on a Yolo V4-tiny convolutional neural network to identify weeds and crops. YOLO V4-tiny is a lightweight convolutional neural network model, which is a simplified version of YOLO V4. Only 600 ten thousand parameters of YOLO V4-tini are one tenth of the parameters of YOLO V4, which greatly improves the detection speed. YOLO V4-tini has the characteristics of multitasking, end-to-end, attention mechanism and multiscale. And the target classification and regression realization parameter sharing can be completed simultaneously by multitasking, so that overfitting is avoided.
Specifically, the steps of establishing the convolutional neural network model to identify weeds and crops and calculating weed density and crop area are as follows.
Step 1: and walking in a field (such as a corn field) by using the mobile platform, acquiring a video, extracting clear pictures from the video, labeling the pictures by using labelimg software, and obtaining a dataset and a corresponding label file after labeling all the pictures by using the labels in two types of Crop and Weed.
Step 2: dividing the data set into a training data set and a testing data set, utilizing the data enhancement expansion data set, combining a migration learning technology to train the YOLO V4-tiny model and optimizing the model weight.
Specifically, the data set manufactured in the step 1 is divided into a training set and a testing set according to the ratio of 8:2, the four pictures are respectively turned over, scaled, color gamut changed and the like by utilizing a mosaic data enhancement algorithm, and the four pictures are well arranged according to the four directions, so that the combination of the pictures and the combination of frames are realized, and the purpose of expanding the data set is further achieved. The YOLO V4-tiny weights trained using the large agricultural dataset are migrated into the YOLO V4-tiny model used herein and hyper-parameters are adjusted to begin training the model.
Step3: and (3) predicting the field weed and crop images by using the model trained in the step (2), comparing the field weed and crop images with the real weed and crop information in the test data set, quantifying the prediction error, and evaluating the accuracy of the model.
Specifically, the model trained in the step 2 is used for predicting the test set, and the performance quality of the model after training is completed is evaluated through comparison of the predicted result and the actual and real result. The Precision and recall Recall, AP, mAP values are important evaluation indexes, and the calculation formula is as follows:
Wherein TP is a positive class, FP is a negative class, FN is a positive class, TN is a negative class.
Where r 1,r2...rn is the Recall value corresponding to the first interpolation of the Precison interpolation segment in ascending order.
All classes of APs are maps, which are formulated as:
by integrating the above evaluation index, the larger the calculated mAP value, the better the model performance is.
Step 4: and (3) identifying the target image acquired in the shading black box by adopting the optimal model evaluated in the step (3), and identifying a weed detection frame and a crop detection frame contained in the target image.
According to the length and width of each identified weed detection frame (generally, a plurality of weed detection frames are contained in the target image) and the coordinates of the central point, the area of each weed detection frame contained in the target image can be calculated, the areas of all the weed detection frames contained in the target image are summed, the overlapping area of the detection frames is removed, the coverage area of weeds in the target image is obtained, and the coverage area of weeds is divided by the area of the target image (namely, the area of a shading black box), so that the weed density of the target area can be calculated.
The area of the crop contained in the target image can be calculated according to the frame length and width of the identified crop detection frame (generally, the target image at most comprises one crop detection frame) and the coordinates of the center point.
Spatially relative terms, such as "upper," "lower," "left," "right," and the like, may be used in the embodiments for ease of description to describe one element or feature's relationship to another element or feature's illustrated in the figures. It will be understood that the spatial terms are intended to encompass different orientations of the device in use or operation in addition to the orientation depicted in the figures. For example, if the device in the figures is turned over, elements described as "under" other elements or features would then be oriented "over" the other elements or features. Thus, the exemplary term "lower" may encompass both an upper and lower orientation. The device may be otherwise positioned (rotated 90 degrees or at other orientations) and the spatially relative descriptors used herein interpreted accordingly.
Moreover, relational terms such as "first" and "second", and the like, may be used solely to distinguish one element from another element having the same name, without necessarily requiring or implying any actual such relationship or order between such elements.
The foregoing has described exemplary embodiments of the invention, it being understood that any simple variations, modifications, or other equivalent arrangements which would not unduly obscure the invention may be made by those skilled in the art without departing from the spirit of the invention.
Claims (9)
1. A flame weeder capable of detecting weed density and crop, characterized in that: comprises a moving platform, a weed density and crop identification device, a leaf gathering mechanism, a fire baffle mechanism, a fire spraying device and a lifting mechanism;
The weed density and crop identification device is positioned at the bottom of the mobile platform and comprises a shading black box, an infrared camera, an LED lamp and an image analysis processing system, wherein the infrared camera, the LED lamp and the image analysis processing system are arranged in the shading black box; two layers of shading curtains are respectively arranged on the front side and the rear side of the shading black box at a certain interval;
the leaf gathering mechanism is arranged at two sides of the rear side outlet of the shading black box and comprises leaf gathering plates and first electric telescopic rods used for controlling the opening angles of the leaf gathering plates, and the opening angles of the leaf gathering plates are controlled through the first electric telescopic rods, so that the space clamped between the two leaf gathering plates is matched with the size of crops;
The fire baffle mechanism is arranged at the rear of the leaf gathering mechanism and comprises two fire baffles and a second electric telescopic rod for controlling the opening angle of the fire baffles, the fire baffles are hinged on the leaf gathering plate through hinges, the second electric telescopic rod is arranged between the fire baffles and the leaf gathering plate, and the bottom end of the fire baffles is provided with an inward bending plate;
The fire spraying device is arranged at two outer sides of the fire baffle mechanism and is connected with a fire spraying control system, and the fire spraying control system adjusts the gas flow according to the density of weeds and the density information of the weeds identified by the crop identification device;
The lifting mechanism is used for controlling the weed density and the overall ground clearance of the crop identification device, the leaf gathering mechanism, the fire baffle mechanism and the fire spraying device, and maintaining the distance from the flame nozzle to the ground within a reasonable range;
In the travelling process of the mobile platform, weeds and crops can lift the window shade to enter/exit the window shade black box, the light shading effect inside the window shade black box is guaranteed under the action of the window shade, and the images of the weeds and the crops passing through the inside of the window shade black box are collected under the auxiliary illumination action of the infrared camera and the LED lamp in the window shade black box; the collected images are input into an image analysis processing system, and the density of weeds and the size information of crops are identified by the image analysis processing system.
2. The flame herbicide capable of detecting weed density and crop as claimed in claim 1, wherein: the leaf gathering mechanism controls the space clamped between the two leaf gathering plates to be matched with the size of the crop according to the density of the weeds and the size information of the crop identified by the crop identification device.
3. The flame herbicide capable of detecting weed density and crop as claimed in claim 1, wherein: and a baffle plate is arranged on the inner side of the leaf gathering plate and used for lifting and gathering the crop leaves, and the angle of the baffle plate is adjusted through a third electric telescopic rod.
4. The flame herbicide capable of detecting weed density and crop as claimed in claim 1, wherein: the flame spraying device comprises a connecting rod and a flame spraying device, and the flame spraying device is arranged on the outer side wall of the flame baffle plate through the connecting rod; the flame thrower comprises a flame shielding cover, a plurality of flame nozzles are arranged on the inner side of the flame shielding cover, and the flame nozzles are connected to a flame spraying control system through pipelines.
5. The flame herbicide capable of detecting weed density and crop as claimed in claim 4, wherein: the flaming control system comprises a gas tank, a gas booster pump and a gas flow controller, wherein the gas inlet of the gas flow controller is connected with the gas outlet of the gas booster pump through a high-pressure hose, and the gas outlet of the gas flow controller is communicated with a flame nozzle of the flaming device through a pipeline.
6. The flame herbicide capable of detecting weed density and crop as claimed in claim 1, wherein: the vertical mounting groove has been seted up on the baffle, and bolt fixed mounting is passed through on the mounting groove at the top of connecting rod, through the mounted position of adjusting the top of connecting rod on the mounting groove, can adjust the initial height of flame thrower apart from ground.
7. The flame herbicide capable of detecting weed density and crop as claimed in claim 1, wherein: the lifting mechanism comprises four longitudinal electric lifting rods which are distributed in a rectangular shape and are arranged at the bottom of the movable platform, and the bottom ends of the four longitudinal electric lifting rods are connected with the shading black box of the weed density and crop identification device.
8. The flame herbicide capable of detecting weed density and crop as claimed in claim 1, wherein: the two sides of the shading black box are provided with deep-sinking wheel supporting mechanisms, each deep-sinking wheel supporting mechanism comprises an elastic telescopic rod and a deep-sinking wheel support, the top ends of the elastic telescopic rods are fixedly arranged on the shading black box, the bottom ends of the elastic telescopic rods are connected with the center of the top of each deep-sinking wheel support, each deep-sinking wheel support is provided with two supporting legs, and the deep-sinking wheels are arranged at the bottoms of the supporting legs; the elastic telescopic rod provides downward elastic restoring force for the deep-sinking wheel support, so that the deep-sinking wheel can be always tightly attached to the ground.
9. Flame weeder capable of detecting weed density and crop according to one of claims 1-8, characterized in that: the image analysis processing system adopts a model based on a YOLO V4-tiny convolutional neural network to identify weeds and crops; the steps of establishing the convolutional neural network model for identifying weeds and crops and calculating weed density and crop area are as follows:
Step 1: collecting infrared images of field weeds and crops through the mobile platform, labeling the images, dividing the labels into crops and weeds, and obtaining a data set and corresponding label files after all the images are labeled;
step 2: dividing the data set into a training data set and a testing data set, utilizing the data enhancement expansion data set, combining a migration learning technology to train the YOLO V4-tiny model and optimizing model weights;
Step 3: predicting field weeds and crop images by using the model trained in the step 2, comparing the field weeds and crop images with real weeds and crop information in the test data set, quantifying prediction errors, and evaluating model accuracy;
Step 4: identifying a target image acquired in the shading black box by adopting the optimal model evaluated in the step 3, and identifying a weed detection frame and a crop detection frame contained in the target image;
Calculating the area of each weed detection frame contained in the target image according to the recognized frame length and width of each weed detection frame and the center point coordinates, summing the areas of all the weed detection frames contained in the target image, removing the overlapping areas of the detection frames, and obtaining the coverage area of weeds in the target image, wherein the coverage area of weeds is divided by the area of the target image, so that the weed density of the target area can be calculated;
And calculating the area of the crops contained in the target image according to the recognized frame length and width of the crop detection frame and the center point coordinates.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211663320.6A CN116114683B (en) | 2022-12-23 | 2022-12-23 | Flame weeding machine capable of detecting weed density and crops |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211663320.6A CN116114683B (en) | 2022-12-23 | 2022-12-23 | Flame weeding machine capable of detecting weed density and crops |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116114683A CN116114683A (en) | 2023-05-16 |
| CN116114683B true CN116114683B (en) | 2024-05-28 |
Family
ID=86303747
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211663320.6A Active CN116114683B (en) | 2022-12-23 | 2022-12-23 | Flame weeding machine capable of detecting weed density and crops |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116114683B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2000250194A (en) * | 1999-03-03 | 2000-09-14 | Lpl:Kk | Black box and black box kit |
| CN109042607A (en) * | 2018-07-10 | 2018-12-21 | 江苏大学 | A kind of the flame weeding machine people system and implementation method of view-based access control model |
| CN110476506A (en) * | 2019-07-30 | 2019-11-22 | 商丘市农林科学院 | The sterilizing pesticide weeding methods of cultivation and equipment |
| CN111813155A (en) * | 2020-06-18 | 2020-10-23 | 五邑大学 | Target spraying system control method and device based on fuzzy PID and storage medium |
| CN112167233A (en) * | 2020-10-23 | 2021-01-05 | 山东农业大学 | Fire spraying weeding robot based on machine vision |
| CN113610040A (en) * | 2021-08-16 | 2021-11-05 | 华南农业大学 | Paddy field weed density real-time statistical method based on improved BiSeNetV2 segmentation network |
| CN114821268A (en) * | 2022-05-06 | 2022-07-29 | 湖州师范学院 | Weed and crop identification method based on machine learning |
-
2022
- 2022-12-23 CN CN202211663320.6A patent/CN116114683B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2000250194A (en) * | 1999-03-03 | 2000-09-14 | Lpl:Kk | Black box and black box kit |
| CN109042607A (en) * | 2018-07-10 | 2018-12-21 | 江苏大学 | A kind of the flame weeding machine people system and implementation method of view-based access control model |
| CN110476506A (en) * | 2019-07-30 | 2019-11-22 | 商丘市农林科学院 | The sterilizing pesticide weeding methods of cultivation and equipment |
| CN111813155A (en) * | 2020-06-18 | 2020-10-23 | 五邑大学 | Target spraying system control method and device based on fuzzy PID and storage medium |
| CN112167233A (en) * | 2020-10-23 | 2021-01-05 | 山东农业大学 | Fire spraying weeding robot based on machine vision |
| CN113610040A (en) * | 2021-08-16 | 2021-11-05 | 华南农业大学 | Paddy field weed density real-time statistical method based on improved BiSeNetV2 segmentation network |
| CN114821268A (en) * | 2022-05-06 | 2022-07-29 | 湖州师范学院 | Weed and crop identification method based on machine learning |
Non-Patent Citations (2)
| Title |
|---|
| 基于视觉的苗期作物株间除草关键技术研究现状;马志艳;朱熠;杨磊;;中国农机化学报;20200215(02);第32页至第38页 * |
| 采用黑舍养鸡技术饲养艾维茵父母代种鸡生产效果观察;罗庆斌, 詹勋, 包世增, 卢志光;养禽与禽病防治;19970905(09);第5页至第7页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116114683A (en) | 2023-05-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11874265B2 (en) | Multi-sensor platform for crop health monitoring | |
| Kirchgessner et al. | The ETH field phenotyping platform FIP: a cable-suspended multi-sensor system | |
| US10372987B2 (en) | Agricultural method and system using a high resolution sensing device for analyzing and servicing crops | |
| US20220065835A1 (en) | Autonomous crop monitoring system and method | |
| CN106441442A (en) | Field crop phenotypic information high-pass peer monitoring device and monitoring method | |
| CN111626985A (en) | Poultry body temperature detection method based on image fusion and poultry house inspection system | |
| CN110487730A (en) | Crop field phenotype high-throughout monitoring system and monitoring method | |
| CN114037552B (en) | Method and system for polling physiological growth information of meat ducks | |
| EP3591838A2 (en) | Vehicle system and method for autonomous and automated maintenance or care of a photovoltaic free surface system | |
| CN114190267A (en) | System for regulating and controlling growth environment of hydroponic plants and acquiring phenotype images | |
| da Costa Lima et al. | Variable rate application of herbicides for weed management in pre-and postemergence | |
| CN116114683B (en) | Flame weeding machine capable of detecting weed density and crops | |
| Khuzaimah et al. | Application and potential of drone technology in oil palm plantation: Potential and limitations | |
| JP2020146030A (en) | Agricultural system | |
| Thompson et al. | Integrating advanced technologies for optimization of aerial herbicide applications | |
| CA3197532A1 (en) | Systems, methods and devices for using machine learning to optimize crop residue management | |
| CN117575835A (en) | Tea garden growth monitoring system and method based on unmanned aerial vehicle aerial photography | |
| CN112580513A (en) | Intelligent identification method for crop diseases and insect pests | |
| Li et al. | Weed detection in soybean fields using improved YOLOv7 and evaluating herbicide reduction efficacy | |
| DE202022102591U1 (en) | System for monitoring plant health in precision agriculture using image processing and convolutional neural network | |
| US20210195850A1 (en) | Greenhouse control system | |
| CN112598531A (en) | Unmanned automatic field patrol method for farm | |
| CN110825118A (en) | Multi-unmanned aerial vehicle cooperative farmland spraying method based on deep learning algorithm | |
| CN218929641U (en) | Vehicle-mounted system for detecting plant diseases and insect pests | |
| KR102433629B1 (en) | Foldable hyperspectral imaging device for phenotype measurement of large-scale cornered crops |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |