Automatic tobacco insect monitoring method and system for tobacco factory
By acquiring images of tobacco beetle traps and performing image preprocessing and neural network recognition, the problem of laborious and error-prone manual identification and counting of tobacco beetles in tobacco factories has been solved, thus improving the accuracy and convenience of tobacco beetle monitoring.
Patent Information
- Application Number
- CN202511147356.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-08-15
- Publication Date
- 2025-11-18
AI Technical Summary
The current practice in tobacco factories of manually identifying and counting tobacco insects by sight is laborious and prone to errors, resulting in poor insect monitoring.
By acquiring images of tobacco beetle traps, preprocessing them using an image processing module, and then using a neural network model to identify and count the number of tobacco beetles, tobacco beetle monitoring data is generated and stored.
This improved the accuracy and convenience of tobacco insect monitoring, and reduced the labor intensity and errors of manual operations.
Smart Images
Figure CN120976865A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of tobacco moth monitoring, and more particularly to a tobacco moth automatic monitoring method and system for a tobacco factory. BACKGROUND
[0002] The existing tobacco factory uses a tobacco beetle trap as a main tool for insect situation monitoring, and a manual visual method is used to identify and count the tobacco moths on the trap after a point is patrolled. When the number of tobacco moths on the trap is large, the counting is prone to error, which cannot provide correct data analysis for insect situation monitoring, thereby resulting in poor insect situation monitoring effect and misjudgment. Therefore, how to automatically monitor the tobacco moths and improve the accuracy and convenience of the tobacco moth monitoring has important significance. SUMMARY
[0003] The present application provides a tobacco moth automatic monitoring method and system for a tobacco factory, which solves the problem of labor and error-prone counting in the existing tobacco factory using a manual visual method to identify and count the tobacco moths, and can improve the accuracy and convenience of the tobacco moth monitoring.
[0004] To achieve the above object, the present application provides the following technical scheme:
[0005] A tobacco moth automatic monitoring method for a tobacco factory, comprising:
[0006] obtaining tobacco beetle trap images of each monitoring point and uploading the tobacco beetle trap images to a server;
[0007] A tobacco beetle image processing module is arranged in the server to pre-process the tobacco beetle trap images.
[0008] A tobacco beetle identification module is further arranged in the server to identify the tobacco moths from the pre-processed tobacco beetle trap images.
[0009] The number of the tobacco moths is counted and analyzed according to the identification result to form and store the tobacco moth monitoring data.
[0010] Preferably, the tobacco beetle trap images of each monitoring point are obtained, comprising:
[0011] The photographing function of a mobile phone of a patrol personnel and a mobile network are used to realize the image acquisition and transmission.
[0012] Preferably, the tobacco beetle trap images of each monitoring point are obtained, further comprising:
[0013] A network monitoring camera is arranged at a set position in front of each tobacco beetle trap to realize the real-time image acquisition and transmission.
[0014] Preferably, the image preprocessing of the smoke beetle trap image includes:
[0015] The initial image is denoised and converted into a grayscale image. The grayscale image is then binarized to enhance the grayscale pixel blocks corresponding to the tobacco insects. After enhancement, the processed image is labeled.
[0016] Preferably, the step of identifying tobacco beetle in the preprocessed image of the tobacco beetle trap includes:
[0017] Build and train a neural network model, and use the trained neural network model to identify tobacco worms.
[0018] Preferably, the construction and training of the neural network model includes:
[0019] Set a convolutional neural network (CNN) as the neural network model;
[0020] Set up a tobacco beetle image database, and construct a training sample set and a test sample set from the data in the tobacco beetle image database according to a certain ratio;
[0021] The training sample set and the test sample set are labeled and then input into the neural network model for training.
[0022] Preferred options also include:
[0023] The tobacco beetle traps are numbered, and the insect population data, abnormal information, risk level, and historical images of the tobacco beetle traps are viewed according to the number.
[0024] Preferred options also include:
[0025] Based on the tobacco insect identification results, abnormal insect infestation is determined. If the number of tobacco insects exceeds the set threshold, an abnormal tobacco insect alarm is reported.
[0026] The present invention also provides an automatic tobacco insect monitoring system for tobacco factories, which uses the above-mentioned monitoring method and includes: a visual acquisition device, an image acquisition terminal, and a server;
[0027] Multiple visual acquisition devices are set up in front of the beetle traps at each monitoring point to take pictures of the beetle traps;
[0028] The image acquisition terminal is signal-connected to the visual acquisition device to acquire images of the beetle traps at each monitoring point and transmit the beetle trap images to the server.
[0029] The server is equipped with a tobacco beetle image processing module and a tobacco beetle recognition module.
[0030] The tobacco beetle image processing module is used to perform image preprocessing on the tobacco beetle trap image.
[0031] The tobacco beetle identification module is used to identify tobacco beetles from the pre-processed image of the tobacco beetle trap.
[0032] The server is equipped with a human-machine interface and performs quantity statistics and analysis on tobacco insects based on the tobacco insect identification results to form and store tobacco insect monitoring data.
[0033] Preferably, the visual acquisition device is a camera, and the image acquisition terminal is an image processing controller.
[0034] This invention provides an automatic monitoring method and system for tobacco beetles in tobacco factories. It acquires images of tobacco beetle traps at various monitoring points, preprocesses these images using a tobacco beetle image processing module, and then identifies the tobacco beetles in the preprocessed images using a tobacco beetle identification module. This generates and stores tobacco beetle monitoring data. This method solves the problem of laborious and error-prone manual visual identification and counting of tobacco beetles in existing tobacco factories, improving the accuracy and convenience of tobacco beetle monitoring. Attached Figure Description
[0035] To more clearly illustrate the specific embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below.
[0036] Figure 1 This is a schematic diagram of an automatic tobacco insect monitoring method for tobacco factories provided by the present invention.
[0037] Figure 2 This is a schematic diagram of a neural network model structure provided in an embodiment of the present invention.
[0038] Figure 3 This is a schematic diagram of an automatic tobacco pest monitoring system for a tobacco factory provided in an embodiment of the present invention. Detailed Implementation
[0039] To enable those skilled in the art to better understand the embodiments of the present invention, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings and implementation methods.
[0040] To address the problems of laborious and error-prone identification and counting of tobacco insects in current methods, this invention provides an automatic tobacco insect monitoring method and system for tobacco factories. This method solves the problem of laborious and error-prone manual visual identification and counting of tobacco insects in existing tobacco factories, and improves the accuracy and convenience of tobacco insect monitoring.
[0041] like Figure 1As shown, an automatic monitoring method for tobacco insects in a tobacco factory includes:
[0042] S1: Acquire images of beetle traps at each monitoring point and transmit the beetle trap images to the server.
[0043] S2: The server is equipped with a tobacco beetle image processing module to preprocess the images of the tobacco beetle trap.
[0044] S3: The server is also equipped with a tobacco beetle identification module to identify tobacco beetles from the pre-processed image of the tobacco beetle trap.
[0045] S4: Based on the tobacco insect identification results, perform quantity statistics and analysis on tobacco insects to form tobacco insect monitoring data and store it.
[0046] Specifically, machine vision acquisition equipment takes daily, timed photos of tobacco beetle traps and transmits the image data to a server. The server converts the captured target's morphological information, pixel distribution, brightness, and color into digital signals, performs various calculations on these signals to extract target features, and then identifies tobacco beetles and detects anomalies based on these extracted features. In practical applications, the tobacco beetle trap can be a paper-based adhesive board containing tobacco beetle pheromones. The visual acquisition equipment captures images of the trap, automatically identifies and counts tobacco beetles, and uploads the beetle count and images to the server for data storage and pest infestation assessment. An alarm is triggered if an anomaly is detected. This method improves the accuracy and convenience of tobacco beetle monitoring.
[0047] Furthermore, images of beetle traps at each monitoring point were acquired, including by using the camera function of the inspectors' mobile phones and mobile networks to collect and transmit images.
[0048] Furthermore, acquiring images of the beetle traps at each monitoring point also includes: setting up network monitoring cameras at designated locations in front of each beetle trap to achieve real-time image acquisition and transmission.
[0049] Furthermore, the image preprocessing of the tobacco beetle trap image includes: denoising the initial image, converting the denoised initial image into a grayscale image, performing binary conversion on the grayscale image to enhance the grayscale pixel blocks corresponding to the tobacco beetles in the image, and labeling the processed image after enhancement.
[0050] In practical applications, data augmentation also involves processing images such as vertical flipping, row and column transposition, random grayscale, rotation, horizontal flipping, grid distortion, Gaussian filtering, random fogging, elasticity adjustment, histogram equalization, and / or mean filtering. After augmentation, the processed data is labeled. Labelimg is a graphical image annotation tool that can be used to label the data. The coordinate information of tobacco insects can be manually labeled and saved in XML format. The XML file contains information such as the location coordinates, species, and image path of the tobacco insects in the image.
[0051] Furthermore, the step of identifying tobacco beetle in the preprocessed image of the tobacco beetle trap includes: constructing and training a neural network model, and identifying tobacco beetles using the trained neural network model.
[0052] Furthermore, the construction and training of the neural network model includes:
[0053] Set a convolutional neural network (CNN) as the neural network model;
[0054] Set up a tobacco beetle image database, and construct a training sample set and a test sample set from the data in the tobacco beetle image database according to a certain ratio;
[0055] The training sample set and the test sample set are labeled and then input into the neural network model for training.
[0056] The method also includes: numbering the tobacco beetle traps, and viewing the insect population data collection records, abnormal information, risk level, and historical tobacco beetle trap images of the corresponding traps according to the numbers.
[0057] In practical applications, QR codes can be used for numbering. Before use, the QR code on the trap should be used to register information such as the placement location, placement time, and personnel involved. The QR code on the trap can be used to track the replacement of the trap and the rate of change in insect population.
[0058] The method also includes: determining abnormal insect infestation based on tobacco insect identification results; and reporting an abnormal tobacco insect alarm if the number of tobacco insects exceeds a set threshold.
[0059] In practical applications, ResNet50 is a residual neural network. This application replaces the repeated convolutions in this neural network with downsampling operations, which have lower computational cost. This reduces the computational cost required for the model to identify tobacco worms while maintaining the quality of the feature maps. The improved network structure is as follows: Figure 2As shown, after constructing the model structure, the model is trained using the labeled training set. The training effect is then evaluated using accuracy, precision, and recall. The model that meets the criteria is selected as the final pest identification model. This model is then used to identify the number of tobacco insects on images of real-time collected traps. After counting the insects, the count is compared with a preset threshold. If the number of insects exceeds the threshold, an alert is sent to relevant quality control personnel via an app. This alert can be sent using widely used software such as email or WeChat.
[0060] Therefore, this invention provides an automatic tobacco beetle monitoring method for tobacco factories. It acquires images of tobacco beetle traps at various monitoring points, preprocesses these images using a tobacco beetle image processing module, and then identifies the tobacco beetles in the preprocessed images using a tobacco beetle identification module to generate and store tobacco beetle monitoring data. This method solves the problem of laborious and error-prone manual visual identification and counting of tobacco beetles in existing tobacco factories, thus improving the accuracy and convenience of tobacco beetle monitoring.
[0061] Accordingly, the present invention also provides an automatic tobacco beetle monitoring system for tobacco factories, using the aforementioned monitoring method, comprising: visual acquisition devices, an image acquisition terminal, and a server. Multiple visual acquisition devices are positioned in front of tobacco beetle traps at each monitoring point to photograph the traps. The image acquisition terminal is signal-connected to the visual acquisition devices to acquire images of the tobacco beetle traps at each monitoring point and transmits these images to the server. The server includes a tobacco beetle image processing module and a tobacco beetle recognition module. The tobacco beetle image processing module performs image preprocessing on the tobacco beetle trap images. The tobacco beetle recognition module identifies the tobacco beetles in the preprocessed images. The server has a human-machine interface and performs quantity statistics and analysis on the tobacco beetles based on the recognition results to generate and store tobacco beetle monitoring data.
[0062] Furthermore, the visual acquisition device is a camera, and the image acquisition terminal is an image processing controller.
[0063] Specifically, such as Figure 3 As shown, the image of the tobacco beetle trap is acquired by a visual acquisition device, and the image acquisition terminal transmits the acquired photos to a computer for inspection. The detection and judgment are performed by the tobacco beetle image processing module and the tobacco beetle recognition module, and the judgment results are output to the statistical analysis module to perform the functions of tobacco beetle infestation early warning and statistical analysis.
[0064] Therefore, this invention provides an automatic tobacco beetle monitoring system for tobacco factories. It acquires images of tobacco beetle traps at various monitoring points using a visual acquisition device, preprocesses these images using a tobacco beetle image processing module, and then identifies the tobacco beetles in the preprocessed images using a tobacco beetle recognition module to generate and store tobacco beetle monitoring data. This system solves the problem of laborious and error-prone manual visual identification and counting of tobacco beetles in existing tobacco factories, thus improving the accuracy and convenience of tobacco beetle monitoring.
[0065] The structure, features, and effects of the present invention have been described in detail above with reference to the embodiments shown in the figures. The above description is only a preferred embodiment of the present invention, but the present invention is not limited to the scope of implementation shown in the figures. Any changes made in accordance with the concept of the present invention, or equivalent embodiments modified to have equivalent changes, shall be within the protection scope of the present invention as long as they do not exceed the spirit covered by the specification and figures.
Claims
1. An automatic monitoring method for tobacco insects in a tobacco factory, characterized in that, include: Acquire images of the beetle traps at each monitoring point and upload the beetle trap images to the server; The server is equipped with a tobacco beetle image processing module to preprocess the images of the tobacco beetle trap. The server is also equipped with a tobacco beetle identification module to identify tobacco beetles from the pre-processed image of the tobacco beetle trap. Based on the tobacco insect identification results, the number of tobacco insects is statistically analyzed to form and store tobacco insect monitoring data.
2. The automatic monitoring method for tobacco weevils in a tobacco factory according to claim 1, characterized in that, Images of beetle traps at each monitoring point were acquired, including: The inspection personnel's mobile phones are used to take photos and transmit images via mobile networks.
3. The automatic monitoring method for tobacco weevils in a tobacco factory according to claim 2, characterized in that, Images of beetle traps at each monitoring point were acquired, including: By placing network monitoring cameras in front of each smoke beetle trap, images can be collected and transmitted in real time.
4. The automatic monitoring method for tobacco weevils in a tobacco factory according to claim 3, characterized in that, The image preprocessing of the smoke beetle trap image includes: The initial image is denoised and converted into a grayscale image. The grayscale image is then binarized to enhance the grayscale pixel blocks corresponding to the tobacco insects. After enhancement, the processed image is labeled.
5. The automatic monitoring method for tobacco weevils in a tobacco factory according to claim 4, characterized in that, The step of identifying tobacco beetles from the preprocessed image of the tobacco beetle trap includes: Build and train a neural network model, and use the trained neural network model to identify tobacco worms.
6. The automatic monitoring method for tobacco weevils in a tobacco factory according to claim 5, characterized in that, The construction and training of the neural network model includes: Set a convolutional neural network (CNN) as the neural network model; Set up a tobacco beetle image database, and construct a training sample set and a test sample set from the data in the tobacco beetle image database according to a certain ratio; The training sample set and the test sample set are labeled and then input into the neural network model for training.
7. The automatic monitoring method for tobacco weevils in a tobacco factory according to claim 6, characterized in that, Also includes: The tobacco beetle traps are numbered, and the insect population data, abnormal information, risk level, and historical images of the tobacco beetle traps are viewed according to the number.
8. The automatic monitoring method for tobacco weevils in a tobacco factory according to claim 7, characterized in that, Also includes: Based on the tobacco insect identification results, abnormal insect infestation is determined. If the number of tobacco insects exceeds the set threshold, an abnormal tobacco insect alarm is reported.
9. An automatic tobacco weevil monitoring system for a tobacco factory, using the monitoring method according to any one of claims 1 to 8, characterized in that, include: Visual acquisition devices, image acquisition terminals, and servers; Multiple visual acquisition devices are set up in front of the beetle traps at each monitoring point to take pictures of the beetle traps; The image acquisition terminal is signal-connected to the visual acquisition device to acquire images of the beetle traps at each monitoring point and transmit the beetle trap images to the server. The server is equipped with a tobacco beetle image processing module and a tobacco beetle recognition module. The tobacco beetle image processing module is used to perform image preprocessing on the tobacco beetle trap image. The tobacco beetle identification module is used to identify tobacco beetles from the pre-processed image of the tobacco beetle trap. The server is equipped with a human-machine interface and performs quantity statistics and analysis on tobacco insects based on the tobacco insect identification results to form and store tobacco insect monitoring data.
10. The automatic tobacco weevil monitoring system for tobacco factories according to claim 9, characterized in that, The visual acquisition device is a camera, and the image acquisition terminal is an image processing controller.