CN118155144A - Vegetable planting pesticide input supervision system and method based on AI vision - Google Patents
Vegetable planting pesticide input supervision system and method based on AI vision Download PDFInfo
- Publication number
- CN118155144A CN118155144A CN202410585120.6A CN202410585120A CN118155144A CN 118155144 A CN118155144 A CN 118155144A CN 202410585120 A CN202410585120 A CN 202410585120A CN 118155144 A CN118155144 A CN 118155144A
- Authority
- CN
- China
- Prior art keywords
- pesticide
- information
- safety
- vegetable
- management information
- 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.)
- Pending
Links
- 239000000575 pesticide Substances 0.000 title claims abstract description 180
- 235000013311 vegetables Nutrition 0.000 title claims abstract description 95
- 238000000034 method Methods 0.000 title claims abstract description 18
- 238000007726 management method Methods 0.000 claims abstract description 94
- 238000011156 evaluation Methods 0.000 claims abstract description 80
- 238000012544 monitoring process Methods 0.000 claims abstract description 66
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 24
- 238000004458 analytical method Methods 0.000 claims abstract description 20
- 238000005457 optimization Methods 0.000 claims abstract description 16
- 238000005516 engineering process Methods 0.000 claims abstract description 10
- 238000001514 detection method Methods 0.000 claims description 19
- 238000012549 training Methods 0.000 claims description 17
- 230000003993 interaction Effects 0.000 claims description 15
- 230000015654 memory Effects 0.000 claims description 11
- 238000012384 transportation and delivery Methods 0.000 claims description 11
- 239000011159 matrix material Substances 0.000 claims description 10
- 238000004590 computer program Methods 0.000 claims description 8
- 241000607479 Yersinia pestis Species 0.000 claims description 7
- 238000009826 distribution Methods 0.000 claims description 5
- 238000003384 imaging method Methods 0.000 claims description 4
- 238000012417 linear regression Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 3
- 238000010219 correlation analysis Methods 0.000 claims description 3
- 238000013439 planning Methods 0.000 claims description 3
- 238000003860 storage Methods 0.000 claims description 3
- 239000003814 drug Substances 0.000 description 9
- 238000010586 diagram Methods 0.000 description 5
- 238000003306 harvesting Methods 0.000 description 5
- 230000006872 improvement Effects 0.000 description 5
- 238000005507 spraying Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 4
- 230000006399 behavior Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000007613 environmental effect Effects 0.000 description 4
- 241000238631 Hexapoda Species 0.000 description 3
- 238000010835 comparative analysis Methods 0.000 description 3
- 229940079593 drug Drugs 0.000 description 3
- 238000002372 labelling Methods 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 239000000447 pesticide residue Substances 0.000 description 3
- 238000013135 deep learning Methods 0.000 description 2
- 235000021384 green leafy vegetables Nutrition 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000013468 resource allocation Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000012098 association analyses Methods 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000003337 fertilizer Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention belongs to the technical field of agriculture, and discloses a vegetable planting pesticide input monitoring system and method based on AI vision, wherein the system comprises the steps of acquiring a real-time monitoring image of a vegetable planting area based on the AI vision technology, and capturing actual input information of pesticide input; counting cosine similarity distance between application state data of the pesticide and actual throwing information through a first algorithm model by the real-time monitoring image, determining pesticide application accuracy of the actual throwing information, and assigning a pesticide safety evaluation coefficient based on the pesticide application accuracy; and selecting the pesticide safety evaluation coefficient corresponding to the land parcel management information in a relative optimization degree, analyzing the correlation degree of the unsafe picking area by means of set analysis, determining the early warning information of the unsafe picking area, and viewing real-time monitoring data, tracing information and early warning information based on a user interface.
Description
Technical Field
The invention relates to the technical field of agriculture, in particular to a vegetable planting pesticide input supervision system and method based on AI vision.
Background
At present, the monitoring of agricultural input products is mainly residual detection at a post-partum stage, and the input products such as pesticides, fertilizers and the like cannot be detected on line in real time; the traceability system can trace the service condition of agricultural input products in the agricultural product production process by inquiring the traceability information, but the traceability information is mainly manually input, so that the traceability information cannot be timely and accurate.
Such as China issued patent: CN109669403B discloses a critical agriculture input article intelligent monitoring system based on DBN-SOFTMAX, which is used for carrying out real-time on-line monitoring on critical agriculture input articles of planting bases, and when the input articles are applied to the planting bases, the varieties of the input articles can be predicted, so that the varieties and the application time of the input articles input by a base administrator in a traceability system are compared.
However, the prior art also has the following drawbacks:
the existing vegetable planting area is not a fixed planting area with a slice shape, an artificial information importing part exists, and pesticide input products are various and have a plurality of circulation links, so that the accuracy of monitoring information corresponding to the vegetable planting area cannot be ensured;
The pesticide input product and the video monitoring can not be completely combined to play a larger role, and the form is a single monitoring mode, so that the timely and accurate tracing information can not be ensured.
In view of the above, the application provides a vegetable planting pesticide input supervision system and method based on AI vision.
Disclosure of Invention
In order to overcome the defects in the prior art, the invention provides a vegetable planting pesticide input supervision system and method based on AI vision.
In order to achieve the above purpose, the present invention provides the following technical solutions: the embodiment of the first aspect of the invention provides an AI vision-based vegetable planting pesticide input supervision system, which comprises an online monitoring module, a traceability evaluation module, a safety early warning module and a man-machine interaction module, wherein the modules are connected in a wired and/or wireless manner to realize data transmission among the modules;
The on-line monitoring module is used for acquiring a real-time monitoring image of the vegetable planting area based on an AI vision technology and capturing actual throwing information of pesticide throwing products; the real-time monitoring image and the actual delivery information are sent to a traceability evaluation module and a man-machine interaction module;
The traceability evaluation module is used for counting cosine similarity distance between the application state data of the pesticide and the actual throwing information through the first algorithm model by the real-time monitoring image, determining the pesticide application accuracy of the actual throwing information, and assigning a pesticide safety evaluation coefficient based on the pesticide application accuracy; the land parcel management information and the corresponding pesticide safety evaluation coefficients are sent to a safety early warning module and a man-machine interaction module;
The safety early warning module is used for selecting the relative optimization degree of the pesticide safety evaluation coefficient corresponding to the land block management information, carrying out set-to-set association degree analysis on the unsafe picking area through set analysis, determining early warning information of the unsafe picking area and sending the early warning information to the man-machine interaction module;
And the man-machine interaction module is used for viewing real-time monitoring data, tracing information and early warning information based on a user interface.
Based on a preferred technical scheme of the invention, the online monitoring module comprises an area management unit and a data recording unit;
The area management unit is used for planning an AI camera arrangement distribution map through tree structure distributed division in the vegetable planting area and storing the planned vegetable planting area in the corresponding data recording unit;
the data recording unit is used for acquiring real-time monitoring images of the corresponding vegetable planting areas based on the AI camera and capturing actual throwing information of pesticide throwing products, wherein the actual throwing information comprises pesticide varieties, application time and application amount; storing the actual delivery information in a blockchain region database;
and associating the data recording unit with the tree structure, and recording node position information corresponding to the tree structure, wherein the node position information is subjected to GPS real-time positioning control.
Based on a preferred technical scheme of the invention, the logic for dividing the vegetable planting area is as follows:
Dividing a vegetable planting area according to administrative areas, wherein the vegetable planting area comprises N pieces of base management information, and each piece of base management information comprises vegetable safety tracing information, planting scale information and information of a corresponding actual manager of a current vegetable planting area;
the planting scale information comprises M pieces of land block management information based on vegetable planting varieties, and each piece of land block management information comprises a planting crop mark, a planting area, a growing condition and a pest damage condition;
And monitoring and managing the M pieces of land block management information to acquire real-time monitoring images, wherein the real-time monitoring images comprise image information with time stamps acquired based on the AI camera.
Based on a preferred technical scheme of the invention, the acquisition logic of the pesticide application accuracy:
taking a real-time monitoring image on the current land parcel management information as an input end of a first algorithm model, and taking a cosine similarity distance between application state data and actual throwing information as a prediction output value;
Setting a target output value for the cosine similarity distance between the application state data and the actual input information by utilizing the learning approximation of the network training result, wherein the target output value is a prediction target of a prediction output value;
taking the difference value between the minimum target output value and the predicted output value as a training target; training the first algorithm model until the sum of prediction accuracy reaches convergence, and stopping training;
The cosine-like distance between the application state data and the actual delivery information is trained and then marked as pesticide application accuracy.
Based on a preferred technical scheme of the invention, the acquisition logic of the pesticide safety evaluation coefficient comprises the following steps:
Acquiring pesticide application accuracy corresponding to M pieces of land management information, dividing evaluation grades based on priori knowledge on the pesticide application accuracy, and carrying out primary assignment on each piece of land management information based on the evaluation grades;
performing secondary assignment on each piece of land block management information;
acquiring pesticide safety evaluation coefficients from the primary assignment and the secondary assignment through a linear regression algorithm; and correlating the pesticide safety evaluation coefficient with the crop mark, and imaging the land mass management information through an image.
Based on a preferred technical scheme of the invention, the selection logic of the relative optimization degree of pesticide safety evaluation is as follows:
Acquiring pesticide safety evaluation coefficients in land parcel management information And the pesticide safety evaluation corresponds to a preset safety detection interval/>; Pesticide safety evaluation coefficient/>And preset safety detection interval/>Comparative analysis,/>Wherein/>Maximum value of safety detection interval,/>A minimum value of the safety detection interval;
Calculating the relative optimization degree of the current pesticide safety evaluation through a formula ；
；
Relative optimality for pesticide safety assessmentSubstituting pesticide safety threshold value corresponding to vegetable safety traceability information；
If it isThe vegetable planting area corresponding to the land block management information is marked as a safe picking area;
If it is The vegetable planting area corresponding to the plot management information is marked as an unsafe picking area.
Based on a preferred technical scheme of the invention, the acquisition logic of the early warning information is as follows:
The result set comprises basic safety, slight early warning, moderate early warning and severe early warning;
Forming a first set pair for each plot management information and the pesticide safety evaluation coefficient corresponding to each plot management information in the unsafe picking area, calculating the association degree between the first set pair and a result set based on a set analysis model, and generating an association degree matrix;
And analyzing the pesticide use safety condition of each plot management information based on the incidence matrix, and determining the early warning information of the unsafe picking area based on the pesticide application safety condition.
The embodiment of the second aspect of the invention provides an AI vision-based vegetable planting pesticide input supervision system, which is based on the realization of the AI vision-based vegetable planting pesticide input supervision method, and comprises the following steps:
Acquiring a real-time monitoring image of a vegetable planting area based on an AI vision technology, and capturing actual throwing information of pesticide throwing products;
counting cosine similarity distance between application state data of the pesticide and actual throwing information through a first algorithm model by the real-time monitoring image, determining pesticide application accuracy of the actual throwing information, and assigning a pesticide safety evaluation coefficient based on the pesticide application accuracy;
selecting the relative optimization degree of the pesticide safety evaluation coefficient corresponding to the land block management information, carrying out set-to-correlation analysis on the unsafe picking area through set analysis, and determining early warning information of the unsafe picking area;
and viewing real-time monitoring data, tracing information and early warning information based on a user interface.
An embodiment of a third aspect of the present invention proposes an electronic device comprising: a processor and a memory, wherein the memory stores a computer program for the processor to call;
The processor executes the vegetable planting pesticide input supervision system based on AI vision by calling a computer program stored in the memory.
An embodiment of a fourth aspect of the present invention proposes a computer-readable storage medium characterized in that: and instructions are stored, and when the instructions run on the computer, the computer is caused to execute the vegetable planting pesticide input supervision system based on AI vision.
The vegetable planting pesticide input supervision system and method based on AI vision has the technical effects and advantages that:
The real-time monitoring and subsequent analysis steps based on the AI vision technology can realize real-time monitoring and quick response, improve pesticide application accuracy, individuation pesticide safety evaluation, optimize resource allocation and early warning mechanism, enhance decision support, and improve user satisfaction and trust, thereby promoting the healthy development of vegetable planting industry and the improvement of social welfare.
Drawings
FIG. 1 is a schematic diagram of a vegetable planting pesticide input supervision system based on AI vision;
FIG. 2 is a diagram of a tree structure distributed partition logic within an on-line monitoring module according to the present invention;
FIG. 3 is a flowchart of a method for supervising the pesticide input into vegetable planting based on AI vision;
Fig. 4 is a schematic structural diagram of an electronic device according to the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
Referring to fig. 1, the present embodiment provides a vegetable planting pesticide input supervision system based on AI vision, which includes an online monitoring module 1, a traceability evaluation module 2, a security evaluation module 3 and a man-machine interaction module 4; all the modules are connected in a wired and/or wireless mode, so that data transmission among the modules is realized;
The on-line monitoring module 1 is used for acquiring a real-time monitoring image of a vegetable planting area based on an AI vision technology and capturing actual throwing information of pesticide throwing products; the real-time monitoring image and the actual delivery information are sent to a traceability evaluation module 2 and a man-machine interaction module 4;
Specifically, as shown in fig. 2, the online monitoring module 1 includes an area management unit 101 and a data recording unit 102;
An area management unit 101 for planning an AI camera arrangement distribution map by means of tree structure distributed division in the vegetable planting area, and storing the planned vegetable planting area in a corresponding data recording unit 102;
The data recording unit 102 is used for acquiring real-time monitoring images of the corresponding vegetable planting areas based on the AI camera and capturing actual throwing information of pesticide throwing products, wherein the actual throwing information comprises pesticide varieties, application time and application amount; storing the actual delivery information in a blockchain region database;
Associating the data recording unit 102 with a tree structure, and recording node position information corresponding to the tree structure, wherein the node position information is controlled by GPS real-time positioning;
The logic for further dividing the vegetable planting area is as follows:
Dividing a vegetable planting area according to administrative areas, wherein the vegetable planting area comprises N pieces of base management information, and each piece of base management information comprises vegetable safety tracing information, planting scale information and information of a corresponding actual manager of a current vegetable planting area;
the planting scale information comprises M pieces of land block management information based on vegetable planting varieties, and each piece of land block management information comprises a planting crop mark, a planting area, a growing condition and a pest damage condition;
And monitoring and managing the M pieces of land block management information to acquire real-time monitoring images, wherein the real-time monitoring images comprise image information with time stamps acquired based on the AI camera.
Further described, the tree structured distributed partitioning logic is:
taking the geographic position of a vegetable planting area as a root node, taking base management information under the vegetable planting area as a first branch node and taking land block management information as a second branch node, wherein: the root node is a father node of the first support node, and the first support node is a father node of the second support node;
Storing the divided vegetable planting areas in corresponding blockchain area databases, associating the blockchain area databases with tree structures, and recording node position information corresponding to the tree structures, wherein the node position information is subjected to GPS real-time positioning control; agricultural products planted in the area A are prevented from being sold outside in the area B to a certain extent.
It should be noted that: through the distributed management mode of the tree structure, the planting conditions and resource distribution of different areas are conveniently known, a user can conveniently inquire, count and manage data according to different levels, the operation conditions, the production capacity and the planting level of the corresponding vegetable planting areas are known, and data are provided for supervision and market investigation.
The traceability evaluation module 2 is used for counting cosine similarity distance between the application state data of the pesticide and the actual throwing information through a first algorithm model, determining the pesticide application accuracy of the actual throwing information, and assigning a pesticide safety evaluation coefficient based on the pesticide application accuracy; the land parcel management information and the corresponding pesticide safety evaluation coefficients are sent to a safety early warning module 3 and a man-machine interaction module 4;
acquisition logic for the pesticide application accuracy:
taking a real-time monitoring image on the current land parcel management information as an input end of a first algorithm model, and taking a cosine similarity distance between application state data and actual throwing information as a prediction output value;
Setting a target output value for the cosine similarity distance between the application state data and the actual input information by utilizing the learning approximation of the network training result, wherein the target output value is a prediction target of a prediction output value;
taking the difference value between the minimum target output value and the predicted output value as a training target; training the first algorithm model until the sum of prediction accuracy reaches convergence, and stopping training;
The cosine-like distance between the application state data and the actual delivery information is trained and then marked as pesticide application accuracy.
It should be noted that: the first algorithm model is based on an image recognition model of a deep learning technology, and can analyze pesticide application states in images and extract application state data. The similarity between the two vectors is measured by calculating the cosine similarity distance between the application state data and the actual throwing information, so that the similarity between the identified application state and the actual throwing pesticide information is measured, the prediction result is optimized by continuously adjusting model parameters, and when the model is trained to a certain degree, the pesticide application accuracy is directly obtained.
The acquisition logic of the pesticide safety evaluation coefficient comprises the following steps:
Acquiring pesticide application accuracy corresponding to M pieces of land management information, dividing evaluation grades based on priori knowledge on the pesticide application accuracy, and carrying out primary assignment on each piece of land management information based on the evaluation grades;
performing secondary assignment on each piece of land block management information;
It should be noted that: the plot management information comprises two-level assignment for each plot management information based on planting area, growth condition and insect pest condition according to different states, but in practice, the two-level assignment is not limited to planting area, growth condition and insect pest condition, and other environmental factors affecting the plants in plot management are taken as an indispensable part in evaluating pesticide safety evaluation of the plants, and the present embodiment is still in a test state, and because of the limitation of objective factors, the effect of planting area, growth condition and insect pest condition on pesticide safety evaluation coefficients is the greatest, so that the plot management information is taken as important influence factors.
Acquiring pesticide safety evaluation coefficients from the primary assignment and the secondary assignment through a linear regression algorithm; and correlating the pesticide safety evaluation coefficient with the crop mark, and imaging the land mass management information through an image.
It should be noted that: and automatically comparing and verifying the application state data of the pesticide input product monitored in real time with the traceability information of the actual input information, and ensuring the non-tamper property of the traceability information by utilizing a block chain technology, so that the credibility of the data is improved.
The safety early warning module 3 is used for selecting the relative optimization degree of the pesticide safety evaluation coefficient corresponding to the land block management information, carrying out set-to-set association degree analysis on the unsafe picking area through set analysis, determining early warning information of the unsafe picking area and sending the early warning information to the man-machine interaction module 4;
It should be noted that: collecting pesticide safety evaluation coefficients corresponding to the management information of each land block; because the pesticide safety evaluation coefficient is assigned based on the pesticide application accuracy in the traceability evaluation module, other data related to pesticide use safety, such as the environmental conditions of plots, crop growth conditions, pesticide residue detection results and the like, are required to be collected, and in order to more accurately reflect the actual pesticide use safety conditions of each plot, the pesticide safety evaluation coefficient is required to be selected in a relatively optimized degree, and the differences among plots can be fully considered, so that the evaluation results are more in line with the actual conditions.
The selection logic of the relative optimization degree of pesticide safety evaluation is as follows:
Acquiring pesticide safety evaluation coefficients in land parcel management information And the pesticide safety evaluation corresponds to a preset safety detection interval/>; Pesticide safety evaluation coefficient/>And preset safety detection interval/>Comparative analysis,/>Wherein/>Maximum value of safety detection interval,/>A minimum value of the safety detection interval;
Calculating the relative optimization degree of the current pesticide safety evaluation through a formula ；
；
Relative optimality for pesticide safety assessmentSubstituting pesticide safety threshold value corresponding to vegetable safety traceability information；
If it isThe vegetable planting area corresponding to the land block management information is marked as a safe picking area;
If it is The vegetable planting area corresponding to the plot management information is marked as an unsafe picking area.
Further description: through the selection of the relative optimization degree, the land plots with higher pesticide use safety risks can be identified, so that the land plots are preferentially subjected to resource allocation and management, the resource utilization efficiency is improved, the safety and quality of vegetable planting are ensured, personalized guidance is provided, continuous improvement is promoted, and the safety and quality of vegetable planting are ensured.
The acquisition logic of the early warning information is as follows:
The result set comprises basic safety, slight early warning, moderate early warning and severe early warning;
Forming a first set pair for each plot management information and the pesticide safety evaluation coefficient corresponding to each plot management information in the unsafe picking area, calculating the association degree between the first set pair and a result set based on a set analysis model, and generating an association degree matrix;
Further, the calculation of the association degree can be based on methods such as fuzzy mathematics, gray association analysis and the like, and a proper algorithm can be selected according to specific situations. The analysis model comprehensively considers a plurality of factors, such as the numerical value of the pesticide safety evaluation coefficient, the change trend, the suitability of the environmental condition of the land parcel and the like. By calculating the association degree between each set pair and the result set, an association degree matrix can be obtained;
And analyzing the pesticide use safety condition of each plot management information based on the incidence matrix, and determining the early warning information of the unsafe picking area based on the pesticide application safety condition.
It should be noted that: according to the incidence matrix, analyzing the pesticide use safety condition of each plot management information, and regarding plots with lower incidence, meaning that the pesticide use has higher risk, the improvement needs to be focused on and taken measures, the supervision and training of the pesticide use can be enhanced, and the pesticide use consciousness and skill of farmers are enhanced; meanwhile, the method can be popularized and used for replacing varieties of pesticides safely and efficiently, and the risk of pesticide residues is reduced.
And feeding back the effect of the implemented optimization measures to the system, and adjusting and optimizing the system according to the actual effect. Through continuous iteration and improvement, the accuracy and the effectiveness of the safety evaluation module are improved, the pesticide use safety condition of the vegetable planting area is comprehensively and objectively evaluated based on set analysis, decision support is provided for a manager, and the safety production of vegetables is promoted.
And the man-machine interaction module 4 is used for viewing real-time monitoring data, tracing information and early warning information based on a user interface.
The vegetable planting pesticide input monitoring system based on AI vision can start from the aspects of real-time online monitoring, information accuracy improvement, intelligent traceability system, early warning and alarm mechanism, user-friendly interface and operation and the like, solves the problems existing in the prior art, and improves the efficiency and accuracy of vegetable planting pesticide input monitoring. The manager can know the planting condition at any time, so that any abnormal or potential problems can be quickly found, intervention and treatment can be timely performed, the problem expansion is prevented, misuse and abuse of pesticides are reduced, the risk of pesticide residues is reduced, the quality and safety of vegetables are improved, on one hand, the trust of consumers on the quality and safety of the vegetables can be enhanced, the brand image and market competitiveness of the vegetables are improved, and the user satisfaction and loyalty are increased; on the other hand, the management system helps managers to formulate different management strategies for different plots, improves the pertinence and effectiveness of management, and more comprehensively knows the overall situation of vegetable planting.
Example 2
Referring to fig. 3, this embodiment, which is not described in detail in the description of embodiment 1, provides a method for supervising vegetable planting pesticide input based on AI vision, comprising the following steps:
Acquiring a real-time monitoring image of a vegetable planting area based on an AI vision technology, and capturing actual throwing information of pesticide throwing products;
The vegetable planting areas are distributed and divided through a tree structure, AI camera arrangement distribution diagrams are planned, and the planned vegetable planting areas are stored in corresponding block chain area databases;
acquiring real-time monitoring images of corresponding vegetable planting areas based on an AI camera, and capturing actual throwing information of pesticide throwing products, wherein the actual throwing information comprises pesticide varieties, application time and application amount; storing the actual delivery information in a blockchain region database;
and associating the data recording unit with the tree structure, and recording node position information corresponding to the tree structure, wherein the node position information is subjected to GPS real-time positioning control.
The logic for dividing the vegetable planting area is as follows:
Dividing a vegetable planting area according to administrative areas, wherein the vegetable planting area comprises N pieces of base management information, and each piece of base management information comprises vegetable safety tracing information, planting scale information and information of a corresponding actual manager of a current vegetable planting area;
the planting scale information comprises M pieces of land block management information based on vegetable planting varieties, and each piece of land block management information comprises a planting crop mark, a planting area, a growing condition and a pest damage condition;
And monitoring and managing the M pieces of land block management information to acquire real-time monitoring images, wherein the real-time monitoring images comprise image information with time stamps acquired based on the AI camera.
Counting cosine similarity distance between application state data of the pesticide and actual throwing information through a first algorithm model by the real-time monitoring image, determining pesticide application accuracy of the actual throwing information, and assigning a pesticide safety evaluation coefficient based on the pesticide application accuracy;
acquisition logic for the pesticide application accuracy:
taking a real-time monitoring image on the current land parcel management information as an input end of a first algorithm model, and taking a cosine similarity distance between application state data and actual throwing information as a prediction output value;
Setting a target output value for the cosine similarity distance between the application state data and the actual input information by utilizing the learning approximation of the network training result, wherein the target output value is a prediction target of a prediction output value;
taking the difference value between the minimum target output value and the predicted output value as a training target; training the first algorithm model until the sum of prediction accuracy reaches convergence, and stopping training;
The cosine-like distance between the application state data and the actual delivery information is trained and then marked as pesticide application accuracy.
The acquisition logic of the pesticide safety evaluation coefficient comprises the following steps:
Acquiring pesticide application accuracy corresponding to M pieces of land management information, dividing evaluation grades based on priori knowledge on the pesticide application accuracy, and carrying out primary assignment on each piece of land management information based on the evaluation grades;
performing secondary assignment on each piece of land block management information;
acquiring pesticide safety evaluation coefficients from the primary assignment and the secondary assignment through a linear regression algorithm; and correlating the pesticide safety evaluation coefficient with the crop mark, and imaging the land mass management information through an image.
The pesticide safety evaluation coefficient corresponding to the land block management information is selected in relative optimization degree, the non-safety picking area is subjected to set-to-correlation analysis through set analysis, the early warning information of the non-safety picking area is determined,
The selection logic of the relative optimization degree of pesticide safety evaluation is as follows:
Acquiring pesticide safety evaluation coefficients in land parcel management information And the pesticide safety evaluation corresponds to a preset safety detection interval/>; Pesticide safety evaluation coefficient/>And preset safety detection interval/>Comparative analysis,/>Wherein/>Maximum value of safety detection interval,/>A minimum value of the safety detection interval;
Calculating the relative optimization degree of the current pesticide safety evaluation through a formula ；
；
Relative optimality for pesticide safety assessmentSubstituting pesticide safety threshold value corresponding to vegetable safety traceability information；
If it isThe vegetable planting area corresponding to the land block management information is marked as a safe picking area;
If it is The vegetable planting area corresponding to the plot management information is marked as an unsafe picking area.
The acquisition logic of the early warning information is as follows:
The result set comprises basic safety, slight early warning, moderate early warning and severe early warning;
Forming a first set pair for each plot management information and the pesticide safety evaluation coefficient corresponding to each plot management information in the unsafe picking area, calculating the association degree between the first set pair and a result set based on a set analysis model, and generating an association degree matrix;
And analyzing the pesticide use safety condition of each plot management information based on the incidence matrix, and determining the early warning information of the unsafe picking area based on the pesticide application safety condition.
And viewing real-time monitoring data, tracing information and early warning information based on a user interface.
Example 3
In this embodiment, the description of embodiment 1 is not described in detail, and a method for supervising the pesticide input into vegetable planting based on AI vision is provided, where the early warning information corresponds to basic safety, slight early warning, moderate early warning and severe early warning in the result set, and the times of early warning with different degrees are respectively displayed through the selection of time periods or regions, and the higher the times, the more serious the illegal phenomenon is indicated, and meanwhile, which regions are frequently illegal can be analyzed.
Specifically, the frequency of dosing is demonstrated by regional planting base plots by dosing frequency analysis.
The safety interval period analysis counts that the base is harvested after the base is subjected to drug spraying for a plurality of days, and a large amount of data monitor that the harvesting frequency of each base is highest after the base is subjected to drug spraying for a plurality of days; whether the safety interval criterion is met.
According to the behavior action and environmental information of the on-site person collected by the camera, category labeling is carried out to form action recognition base map materials, such as a medicine taking behavior base map labeling person and a medicine taking barrel, and a behavior base map labeling person, a picking basket and vegetables are picked; the auxiliary labels are green leaf vegetables and non-green leaf vegetables, and the vegetables are picked from the existence to the non-existence. Through a large number of base map labels, an action recognition base map model library is established, and objects in images, including people, crops, sprayers and the like, can be recognized by an object detection algorithm based on deep learning. And then judging whether the crops are dosed or not by analyzing the relative positions and actions of the human and the sprayer.
Combining the pesticide spraying time and the harvesting time on the land block management information, comparing the time interval between the harvesting time and the last pesticide spraying time with a safety interval period threshold value, and automatically generating early warning records of different grades;
Marking the harvesting time as a, marking the last time of drug spraying as b, marking the safety interval period threshold as c, wherein the upper limit value of the safety interval period threshold is marked as c1, and the upper limit value of the safety interval period threshold is marked as c2;
Judging early warning information according to the value of c1< (a-b) < c2, wherein the early warning information comprises mild early warning, moderate early warning and severe early warning, the moderate early warning and the severe early warning are marked as violations, and the current land block is subjected to the condition that the harvesting time is standard and the rule breaking frequency is +1;
Specific examples are when the time interval of the time of picking and the time of last time of taking a medicine is a weight early warning in 3-5 days, when the time interval of the time of picking and the time of last time of taking a medicine is a moderate early warning in 6-8 days, when the time interval of picking and the time of last time of taking a medicine is a light early warning in 9-15 days, when the time interval of picking and the time of last time of taking a medicine is more than 15 days, it is normal.
The algorithm model is issued to an AI camera, the AI camera acquires real-time pictures, and the real-time pictures are compared with a model library to feed back the monitoring result of the behavior.
Example 4
An electronic device is shown according to an exemplary embodiment, comprising: a processor and a memory, wherein the memory stores a computer program for the processor to call;
The processor executes the vegetable planting pesticide input supervision system based on AI vision by calling the computer program stored in the memory.
Fig. 4 is a schematic structural diagram of an electronic device provided in an embodiment of the present application, where the electronic device may have a relatively large difference due to different configurations or performances, and may include one or more processors (Central Processing Units, CPU) and one or more memories, where at least one computer program is stored in the memories, and the at least one computer program is loaded and executed by the processors to implement the AI vision-based vegetable planting pesticide input supervision system provided in the above method embodiments.
The electronic device can also include other components for implementing the functions of the device, for example, the electronic device can also have a wired or wireless network interface, an input-output interface, and the like, for input-output. The embodiments of the present application are not described herein.
A computer-readable storage medium, characterized by: and instructions are stored, and when the instructions run on the computer, the computer is caused to execute the vegetable planting pesticide input supervision system based on AI vision.
Those of ordinary skill in the art will appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, or combinations of computer software and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, and are not repeated herein.
The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Finally: the foregoing description of the preferred embodiments of the invention is not intended to limit the invention to the precise form disclosed, and any such modifications, equivalents, and alternatives falling within the spirit and principles of the invention are intended to be included within the scope of the invention.
Claims (10)
1. The vegetable planting pesticide input supervision system based on AI vision is characterized by comprising an online monitoring module (1), a traceability evaluation module (2), a safety early warning module (3) and a man-machine interaction module (4), wherein the modules are connected in a wired and/or wireless mode to realize data transmission among the modules;
the on-line monitoring module (1) is used for acquiring real-time monitoring images of vegetable planting areas based on an AI vision technology and capturing actual throwing information of pesticide throwing products; the real-time monitoring image and the actual delivery information are sent to a traceability evaluation module (2) and a man-machine interaction module (4);
The traceability evaluation module (2) is used for counting cosine similarity distance between the application state data of the pesticide and the actual throwing information through the first algorithm model by the real-time monitoring image, determining the pesticide application accuracy of the actual throwing information, and assigning a pesticide safety evaluation coefficient based on the pesticide application accuracy; the land parcel management information and the corresponding pesticide safety evaluation coefficients are sent to a safety early warning module (3) and a man-machine interaction module (4);
The safety early warning module (3) is used for selecting the relative optimization degree of the pesticide safety evaluation coefficient corresponding to the land block management information, carrying out set-to-set association degree analysis on the unsafe picking area through set analysis, determining early warning information of the unsafe picking area and sending the early warning information to the man-machine interaction module (4);
and the man-machine interaction module (4) is used for viewing real-time monitoring data, tracing information and early warning information based on a user interface.
2. The AI vision-based vegetable-planting pesticide input supervision system according to claim 1, wherein the online monitoring module (1) comprises an area management unit (101) and a data recording unit (102);
The area management unit (101) is used for planning an AI camera arrangement distribution map through tree structure distributed division in the vegetable planting area, and storing the planned vegetable planting area in the corresponding data recording unit (102);
the data recording unit (102) is used for acquiring real-time monitoring images of the corresponding vegetable planting areas based on the AI camera and capturing actual throwing information of pesticide throwing products, wherein the actual throwing information comprises pesticide varieties, application time and application amount; storing the actual delivery information in a blockchain region database;
and associating the data recording unit (102) with the tree structure, and recording node position information corresponding to the tree structure, wherein the node position information is controlled by GPS real-time positioning.
3. The AI vision-based vegetable-planting pesticide input supervision system of claim 2, wherein the logic of dividing the vegetable-planting area is:
Dividing a vegetable planting area according to administrative areas, wherein the vegetable planting area comprises N pieces of base management information, and each piece of base management information comprises vegetable safety tracing information, planting scale information and information of a corresponding actual manager of a current vegetable planting area;
the planting scale information comprises M pieces of land block management information based on vegetable planting varieties, and each piece of land block management information comprises a planting crop mark, a planting area, a growing condition and a pest damage condition;
And monitoring and managing the M pieces of land block management information to acquire real-time monitoring images, wherein the real-time monitoring images comprise image information with time stamps acquired based on the AI camera.
4. The AI vision-based vegetable-planting pesticide input supervision system as set forth in claim 3, wherein: acquisition logic for the pesticide application accuracy:
taking a real-time monitoring image on the current land parcel management information as an input end of a first algorithm model, and taking a cosine similarity distance between application state data and actual throwing information as a prediction output value;
Setting a target output value for the cosine similarity distance between the application state data and the actual input information by utilizing the learning approximation of the network training result, wherein the target output value is a prediction target of a prediction output value;
taking the difference value between the minimum target output value and the predicted output value as a training target; training the first algorithm model until the sum of prediction accuracy reaches convergence, and stopping training;
The cosine-like distance between the application state data and the actual delivery information is trained and then marked as pesticide application accuracy.
5. The AI vision-based vegetable-planting pesticide input supervision system of claim 4, wherein the pesticide safety evaluation coefficient acquisition logic:
Acquiring pesticide application accuracy corresponding to M pieces of land management information, dividing evaluation grades based on priori knowledge on the pesticide application accuracy, and carrying out primary assignment on each piece of land management information based on the evaluation grades;
performing secondary assignment on each piece of land block management information;
acquiring pesticide safety evaluation coefficients from the primary assignment and the secondary assignment through a linear regression algorithm; and correlating the pesticide safety evaluation coefficient with the crop mark, and imaging the land mass management information through an image.
6. The AI vision-based vegetable-planting pesticide input supervision system of claim 5, wherein the selection logic of the relative optimality of pesticide safety assessment is:
Acquiring pesticide safety evaluation coefficients in land parcel management information And the pesticide safety evaluation corresponds to a preset safety detection interval/>; Pesticide safety evaluation coefficient/>And preset safety detection interval/>The comparison and analysis were carried out on the samples,Wherein/>Maximum value of safety detection interval,/>A minimum value of the safety detection interval;
Calculating the relative optimization degree of the current pesticide safety evaluation through a formula ；
；
Relative optimality for pesticide safety assessmentSubstituting pesticide safety threshold value/>, corresponding to vegetable safety traceability information；
If it isThe vegetable planting area corresponding to the land block management information is marked as a safe picking area;
If it is The vegetable planting area corresponding to the plot management information is marked as an unsafe picking area.
7. The AI vision-based vegetable-planting pesticide input supervision system of claim 4, wherein the acquisition logic of the early warning information is:
The result set comprises basic safety, slight early warning, moderate early warning and severe early warning;
Forming a first set pair for each plot management information and the pesticide safety evaluation coefficient corresponding to each plot management information in the unsafe picking area, calculating the association degree between the first set pair and a result set based on a set analysis model, and generating an association degree matrix;
And analyzing the pesticide use safety condition of each plot management information based on the incidence matrix, and determining the early warning information of the unsafe picking area based on the pesticide application safety condition.
8. The AI vision-based vegetable planting pesticide input supervision method is characterized by comprising the following steps of:
Acquiring a real-time monitoring image of a vegetable planting area based on an AI vision technology, and capturing actual throwing information of pesticide throwing products;
counting cosine similarity distance between application state data of the pesticide and actual throwing information through a first algorithm model by the real-time monitoring image, determining pesticide application accuracy of the actual throwing information, and assigning a pesticide safety evaluation coefficient based on the pesticide application accuracy;
the pesticide safety evaluation coefficient corresponding to the land block management information is selected in relative optimization degree, the non-safety picking area is subjected to set-to-correlation analysis through set analysis, the early warning information of the non-safety picking area is determined,
And viewing real-time monitoring data, tracing information and early warning information based on a user interface.
9. An electronic device, characterized in that: comprising the following steps: a processor and a memory, wherein the memory stores a computer program for the processor to call;
The processor executes the AI vision-based vegetable-planting pesticide-input supervision system of any one of claims 1 to 7 by calling a computer program stored in the memory.
10. A computer-readable storage medium, characterized by: instructions stored which, when run on a computer, cause the computer to perform the AI vision-based vegetable-growing pesticide input supervision system as set forth in any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410585120.6A CN118155144A (en) | 2024-05-13 | 2024-05-13 | Vegetable planting pesticide input supervision system and method based on AI vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410585120.6A CN118155144A (en) | 2024-05-13 | 2024-05-13 | Vegetable planting pesticide input supervision system and method based on AI vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN118155144A true CN118155144A (en) | 2024-06-07 |
Family
ID=91285526
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410585120.6A Pending CN118155144A (en) | 2024-05-13 | 2024-05-13 | Vegetable planting pesticide input supervision system and method based on AI vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118155144A (en) |
-
2024
- 2024-05-13 CN CN202410585120.6A patent/CN118155144A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111582055B (en) | Unmanned aerial vehicle aviation pesticide application route generation method and system | |
| CN109840549B (en) | Method and device for identifying plant diseases and insect pests | |
| US20200349477A1 (en) | Closed Loop Integrated Pest Management | |
| Arakeri et al. | Computer vision based robotic weed control system for precision agriculture | |
| EP3482630B1 (en) | Method, system and computer program for performing a pest forecast | |
| CN109146130A (en) | Agricultural product customization is planted and whole process is traced to the source platform and method | |
| CN111476149A (en) | Plant cultivation control method and system | |
| US20220272907A1 (en) | Automated plant monitoring systems and methods | |
| US20220414795A1 (en) | Crop disease prediction and treatment based on artificial intelligence (ai) and machine learning (ml) models | |
| CN108829762A (en) | The Small object recognition methods of view-based access control model and device | |
| KR20210059561A (en) | System and method for pest management of win-win type | |
| CN110057764B (en) | Pesticide application safety management warning device and method | |
| CN117057946A (en) | Intelligent agricultural park data monitoring device and method | |
| CN116543347A (en) | Intelligent insect condition on-line monitoring system, method, device and medium | |
| CN118155144A (en) | Vegetable planting pesticide input supervision system and method based on AI vision | |
| CN115468598A (en) | Intelligent monitoring method and system for pigsty environment | |
| CN113408334B (en) | Crayfish full-chain data acquisition and intelligent detection method and device | |
| US20240016136A1 (en) | Method and system for monitoring and controlling the presence of at least one type of insect in agricultural crops | |
| WO2022067420A1 (en) | Platform for real-time identification and resolution of spatial production anomalies in agriculture | |
| Malik et al. | Machine learning-based Prediction of Cotton farming using ARIMA Model | |
| De Baerdemaeker et al. | Good Agricultural Practices, Quality, Traceability, and Precision Agriculture | |
| KR102271864B1 (en) | Smart farm environment control system that can provide automatic cultivation data based on learning data | |
| JP2024007918A (en) | Advertisement providing system | |
| CN118011970A (en) | Intelligent agricultural production management system based on Internet of things | |
| Senanayaka et al. | Surveillance and Predictive Information System for Tea Smallholdings (SPIS-TS) |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination |