Device based on vector biological database and application method

Abstract

The invention provides equipment based on a vector biological database and an application method, the equipment comprises a fixing assembly and an acquisition terminal, the fixing assembly comprises a storage table, a control support is arranged on the storage table, the acquisition terminal is arranged on the control support, and the acquisition terminal is provided with an identification module, a marking module, a counting module and a classification module; wherein a sample is firstly placed on the object placing table, then the position of the collection terminal is controlled through the control support so that the collection terminal can sample the sample, and finally information obtained through sampling is recognized through the recognition module, marked through the marking module and classified through the classification module. According to the invention, sample collection, image acquisition, intelligent identification, marking, accurate counting and classification can be realized, and the whole process is automatic and intelligent, so that the working time for screening the vector organisms from the large-flux sample is shortened; therefore, the timeliness, accuracy and intelligent level of public health monitoring, port import and export cargo customs clearance and vector biological prevention and control can be greatly improved.

Description

Technical Field

[0001] This invention relates to the field of vector-borne disease technology, and in particular to a device and application method based on a vector-borne disease database. Background Technology

[0002] Vector-borne organisms, such as mosquitoes, flies, rats, ticks, and cockroaches, are key vectors for the transmission of many major infectious diseases, including dengue fever, malaria, plague, and Lyme disease. Continuous, accurate, and efficient monitoring and population dynamic analysis of these organisms are the cornerstone of epidemic early warning and scientific prevention and control in modern public health systems. Traditional vector-borne organism monitoring workflows are highly reliant on manual labor, primarily involving: on-site deployment of traps (such as mosquito lamps and sticky rat traps) → periodic sample collection → manual identification, classification, and counting by professionals under a microscope in the laboratory based on morphological characteristics → manual data recording and reporting. This traditional model exposes the following inherent bottlenecks and serious challenges when addressing increasingly complex public health challenges:

[0003] 1. High reliance on specialized expertise and low efficiency: Accurate identification of specific species requires specialized technicians with extensive training. The manual identification process is time-consuming and labor-intensive, with low sample processing throughput. When facing large-scale monitoring or emergency response to epidemics, human resources become a key constraint, resulting in long monitoring cycles and slow data output.

[0004] 2. Subjective error and consistency challenges: Identification results are easily affected by subjective factors such as personnel experience and fatigue. Different personnel, or even the same person, may make different judgments on the same batch of samples at different times, making it difficult to guarantee the consistency and objectivity of the data and affecting the reliability of long-term trend analysis.

[0005] 4. Counting and statistics are tedious and prone to errors: For high-density samples (such as a large number of mosquitoes collected by mosquito-attracting lamps), manual counting is not only a huge workload, but also prone to statistical errors due to visual fatigue or individual sticking together, affecting the accuracy of population density, a key indicator.

[0006] In recent years, with the rapid development of computer vision and artificial intelligence technologies, especially the breakthroughs in deep learning in the field of general object recognition and classification, new technological paths have been provided for solving the aforementioned problems. Some general-purpose image recognition applications and automated counting devices for agriculture and industry have emerged on the market. However, directly applying these technologies to the professional field of vector-borne disease monitoring still has significant limitations:

[0007] 1. Lack of domain specificity: The general identification model lacks a dedicated database optimized for the fine morphological characteristics of disease vectors (such as wing veins, body bristles, mouthpart structure, etc.), making it difficult to achieve high-precision "genus-species" level classification and failing to meet the stringent requirements for accurate identification of vector species in the tracing of infectious disease sources.

[0008] 2. Insufficient scene adaptability: The accuracy of general object detection and counting algorithms drops sharply in this chaotic background, and it is easy to miss detections, false detections and segmentation errors.

[0009] 3. Incomplete Business Loop: Existing technologies are mostly single-function software tools or algorithm modules, failing to deeply integrate with vector surveillance, standardized business processes, and public health data management systems to form an end-to-end intelligent solution. Identification results cannot be automatically bound to spatiotemporal information, making direct integration with existing vector-borne disease monitoring information management systems difficult. Summary of the Invention

[0010] The purpose of this invention is to identify, label, count, and classify collected samples based on a vector-borne disease database. Specific implementation methods include: 1. Solidifying the shooting background to reduce the sharp drop in accuracy during computation, which could lead to missed detections, false detections, and segmentation errors; 2. Deeply integrating vector-borne disease expertise, core artificial intelligence algorithms, and IoT hardware technology with a professional vector-borne disease image feature database as the core driver, achieving full automation and intelligence from on-site sample collection, automatic image acquisition, intelligent identification and classification, accurate counting, and structured data labeling and uploading. This shortens the time required for screening vector-borne diseases in high-throughput samples, significantly improving the timeliness, accuracy, and intelligence of public health monitoring.

[0011] To achieve the above objectives, the present invention provides the following technical solution:

[0012] Firstly, a device based on a vector-borne organism database includes:

[0013] A fixed component includes a shelf on which a control bracket is provided;

[0014] A data acquisition terminal is mounted on the control bracket, and the data acquisition terminal has an identification module, a marking module, a counting module, and a classification module;

[0015] The process involves first placing the sample on the platform, then controlling the position of the acquisition terminal via the control bracket to sample the sample, and finally identifying the sample through the identification module, marking it through the tagging module, and classifying it through the classification module.

[0016] As described above, the device based on a vector-borne organism database further includes a clamping bracket comprising a telescopic robotic arm and a clamping component. The clamping component is disposed at one end of the telescopic robotic arm and consists of two parts: one part has a gear pin and the other part has a serrated groove. The gear pin is adapted to the serrated groove.

[0017] A telescopic support rod is installed inside the fixed bracket. One end of the telescopic support rod is connected to the fixed bracket via a spherical structural member. A threaded fastener is also installed inside the fixed bracket. The surface of the telescopic support rod has textures to fit the threaded fastener. The threaded fastener is also equipped with a screw knob. Rotating the screw knob controls the tightness between the telescopic support rod and the threaded fastener. A plastic gasket is installed at the other end of the telescopic support rod.

[0018] In the device based on the vector-borne organism database described above, the bottom of the platform is provided with a sample tray, one end of the sample tray is provided with a pull ring, the two sides of the other end of the sample tray are inclined upwards, and an opening is provided in the middle of the other end of the sample tray.

[0019] In the device based on the vector-borne organism database described above, the inner wall of the platform is further provided with a surround-type variable temperature lamp, which includes an inner light strip and an outer soft light cover, the outer soft light cover being made of a semi-transparent frosted material.

[0020] As described above, the device based on the vector-borne organism database further includes a groove on the inner wall of the support connecting the platform and the control bracket. An embedded light is installed in the groove. The embedded light includes an inner light source and an outer light cover. The embedded light can be popped out or fixed by pressing, and the embedded light has a foldable and rotatable bracket.

[0021] In the device based on the vector-borne organism database described above, the gear pin is cross-shaped, the thickness of the tooth tip circle of the gear pin is 0.1 mm, and the spacing between each tooth of the serrated groove is 0.15 mm.

[0022] In the vector-borne organism database-based device described above, the data collection terminal is a mobile phone with a camera function, and the mobile phone has an application program installed that includes the identification module, the tagging module, the counting module, and the classification module.

[0023] Secondly, a device application method based on a vector-borne organism database, using the aforementioned device, includes the following specific steps:

[0024] Create a task queue;

[0025] Obtain sample image information, identify the sample image information through a preset algorithm service layer, and put it into the task queue;

[0026] Preprocess the sample image information in the task queue;

[0027] Build a deep learning model, connect the deep learning model with a vector-borne organism database, and train it.

[0028] The pre-processed sample image information is analyzed and compared using a trained deep learning model;

[0029] The analysis and comparison results are used to generate a report, and key information is marked in the report.

[0030] The device application method based on the vector-borne organism database described above further includes the following specific steps after preprocessing the sample image information in the task queue:

[0031] Construct a YoLo11n model;

[0032] Load the weights from YoLon.pt to fine-tune the YoLo11n model;

[0033] The fine-tuned YoLo11n model was tested in conjunction with the built deep learning detection framework;

[0034] The tested YoLo11n model is imported into the pre-processed sample image information for counting;

[0035] Generate a report from the counting results.

[0036] The device application method based on the vector-borne organism database described above further includes the following specific steps after preprocessing the sample image information in the task queue:

[0037] Construct a multi-class image classification model;

[0038] The multi-class image classification model was pre-trained on the ImageNet dataset.

[0039] The pre-trained model is imported into the pre-processed sample image information for classification.

[0040] The classification results were validated using the cross-entropy loss function and the AdamW optimization algorithm.

[0041] Generate a report from the verified classification results.

[0042] Compared with the prior art, the advantages of this invention are as follows:

[0043] This invention reduces the risk of missed detections, false detections, and segmentation errors caused by a sharp drop in accuracy during computation by fixing the shooting background. Driven by a vector-borne biological image feature database, it can automate and intelligently complete the entire process from on-site sample collection, image acquisition, intelligent identification, labeling, accurate counting, and classification. This shortens the working time for screening vector-borne organisms from high-throughput samples, and can significantly improve the timeliness, accuracy, and intelligence of public health monitoring, customs clearance of import and export goods at ports, and vector-borne biological control. Attached Figure Description

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0045] Figure 1 This is a schematic diagram of the overall device in an embodiment of the present invention;

[0046] Figure 2 This is a structural exploded view of the shelf in an embodiment of the present invention;

[0047] Figure 3 This is a schematic diagram of the sample tray structure in an embodiment of the present invention;

[0048] Figure 4 This is a structural exploded view of the clamping component in an embodiment of the present invention;

[0049] Figure 5 This is a structural exploded view of the telescopic support rod in an embodiment of the present invention;

[0050] Figure 6 This is a flowchart illustrating the principle of the application method in this embodiment of the invention;

[0051] In the diagram: 1. Data acquisition terminal; 2. Fixing component; 3. Control bracket; 4. Display platform; 5. Clamping bracket; 6. Telescopic support rod; 7. Fixing bracket; 8. Recessed light; 9. Groove; 10. Application program; 11. Clamping component; 12. Telescopic robotic arm; 13. Surround temperature-changing lamp; 14. Control knob; 15. Sample tray; 16. Outer diffuser; 17. Inner light strip; 18. Serrated groove; 19. Gear pin; 20. Screw knob; 21. Threaded fastener; 22. Spherical structural component; 23. Texture; 24. Plastic gasket; 25. Inner light source; 26. Outer lamp cover. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0053] Example:

[0054] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, in the embodiments of this invention are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0055] It should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," "counterclockwise," "axial," "radial," and "circumferential" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention.

[0056] In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified. Furthermore, unless otherwise explicitly specified and limited, the terms "installed," "connected," and "linked" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0057] In this invention, unless otherwise explicitly specified and limited, "above" or "below" the second feature can mean that the first feature is in direct contact with the second feature, or that the first feature is in indirect contact with the second feature through an intermediate medium. Furthermore, "above," "over," and "on top" of the second feature can mean that the first feature is directly above or diagonally above the second feature, or simply that the first feature is at a higher horizontal level than the second feature. "Below," "below," and "under" the second feature can mean that the first feature is directly below or diagonally below the second feature, or simply that the first feature is at a lower horizontal level than the second feature.

[0058] In a first aspect, the present invention provides a technical solution: a device based on a vector-borne organism database, see [link to relevant documentation]. Figures 1 to 5 The system includes a sample collection terminal 1, preferably a mobile phone with a camera. Compared to other large devices, mobile phones are lighter and more portable. An application 10 is installed on the mobile phone, which includes an identification module, a marking module, a counting module, and a classification module to identify, mark, count, and classify sample images captured by the mobile phone 10. By connecting to a vector-borne disease database, vector-borne disease samples in the photos are identified, and framed, counted, and labeled. Non-vector-borne disease samples are not labeled. By comparing the photos with labeled vector-borne disease samples to the actual sample locations, vector-borne disease samples can be picked out more quickly.

[0059] In this invention, the fixing component 2 includes a platform 4 and a control bracket 3; the platform 4 includes a sample tray 15, a surround-type variable temperature lamp 13, and an embedded lamp 8. The sample tray 15 is located at the bottom of the platform 4 for placing the collected samples; the front end of the sample tray 15 has a pull ring, and both sides of the end of the sample tray 15 are raised upwards by 10°. The middle of the end of the sample tray 15 has an opening for pouring and cleaning the samples. The sample tray 15 is white to provide a uniform background for the collected samples to improve sample identification; the surround-type variable temperature lamp 13 includes an inner light strip 17 and an outer diffuser 16. The outer diffuser 16 is used to convert direct light into diffused light, eliminating harsh shadows and highlights. The ambient temperature lamp 13 is set around the inner side wall of the platform 4 to illuminate the sample placed on the sample tray 15. The outer diffuser 16 is made of a semi-transparent frosted material with a thickness of 0.2mm. The ambient temperature lamp 13 adjusts the color temperature and brightness through the control knob 14. The recessed lamp 8 includes an inner light source 26 and an outer lamp cover 25. The recessed lamp 8 is set in the inner wall of the bracket connecting the platform 4 and the control bracket 3. The recessed lamp 8 can be removed from the groove 9 in the inner wall of the bracket. After being removed, the recessed lamp 8 is turned on. After being pressed back into the groove 9 in the inner wall of the bracket, the recessed lamp 8 is turned off. The recessed lamp 8 has a foldable and rotatable bracket to provide auxiliary light to compensate for the lack of light in the shadow areas illuminated by the ambient temperature lamp 13.

[0060] In this invention, the control bracket 3 includes a clamping bracket 5 and a fixing bracket 7. The clamping bracket 5 includes a telescopic robotic arm 12 and a clamping component 11. The telescopic robotic arm 12 is used to adjust the shooting position of the mobile phone. The clamping component 11 is divided into left and right parts. The left half of the clamping component 11 has four rows of gear pins 19, and the right half of the clamping component 11 has three rows of serrated grooves 18. The fixed bracket 7 has a telescopic support rod 6 inside. One end of the telescopic support rod 6 is equipped with a ball-shaped structural member 22 to fix the telescopic support rod 6 and prevent it from slipping out of the fixed bracket 7. The fixed bracket 7 also has a threaded fastener 21 to fix the telescopic support rod 6. The threaded fastener 21 is equipped with a screw knob 20. Rotating the screw knob 20 to the left can fix the threaded fastener 21 and the telescopic support rod 6. Rotating the screw knob 20 to the right can loosen the threaded fastener 21 and the telescopic support rod 6, allowing the telescopic support rod 6 to extend and retract freely. The other end of the telescopic support rod 6 is equipped with a plastic gasket 24 to fix the upper part of the mobile phone. The surface of the telescopic support rod 6 has a texture 23. The texture material is a small metal particle frosted material to increase the friction with the threaded fastener 21 and improve anti-slip and stability.

[0061] In this invention, the gear pin 19 is cross-shaped, and the thickness of the tooth tip circle of the gear pin 19 is 0.1mm. It is used to be vertically fixed in the serrated groove 18. The interval between each serration of the serrated groove 18 is 0.15mm. The clamping member 11 is set in the lower half of the platform 4 to avoid the volume control keys and power button on the left and right sides of the mobile phone. The clamping member 11 can adjust and fix the width of the mobile phone according to different models of mobile phones.

[0062] Secondly, the present invention also provides another technical solution: a device application method based on a vector-borne organism database, which uses the above-mentioned device and connects to the vector-borne organism database to identify, mark, count and classify.

[0063] The labeling module primarily relies on the recognition module and the Celery task queue. When the AI ​​algorithm service layer receives a recognition task, it places it in a shared message queue. Multiple pre-started CeleryWorkers, each equipped with a trained deep learning model, evenly distribute the incoming tasks. Each worker, after receiving an image, preprocesses it, such as resizing, noise reduction, and format conversion, to meet the model's analysis standards. The processed image is then fed into the deep learning model for deep analysis. The model extracts key morphological features of the vector organisms and performs rapid comparisons and complex mathematical operations with data from the vector organism database. After independently completing the analysis, each worker generates a report containing the three most likely candidate species and their credibility, among other key information. This report is then sent back to the service layer, which integrates the recognition results, generates a report, and displays it to the user with bounding boxes. Credibility measures the accuracy and reliability of the samples; its core is the quantification of the reliability of the estimation results using tools such as confidence intervals.

[0064] The counting module can use the YoLo11n model. First, install ultralytics (pip install ultralytics) as the library for the YoLo11n model. Then, build a deep learning detection framework on Ubuntu 20.0.4, using Python 3.8 and PyTorch 1.9.1 as the training and testing environment. Load the pre-trained YoLon.pt weights to fine-tune the YoLo11n model. This model has two input image sizes: 640×640 pixels and 1024×1024 pixels, corresponding to a lightweight small-size model (best_0.9048_640.pt) and a lightweight high-precision model (best_0.9520_1024.pt), respectively. Each model is trained for 300 epochs. The SGD optimizer is selected to optimize the network model parameters. The initial learning rate is set to 0.001 to achieve network model convergence. After that, the fine-tuned model is tested in conjunction with the built deep learning detection framework. Finally, the tested model is imported into the pre-processed sample image information for counting. The program integrates the average rate of each identified category and the number of effective samples to generate a report for display to the user.

[0065] The classification module, based on samples from a vector-borne disease database, utilizes the PyTorch deep learning framework to construct a multi-class image classification model training and validation process, thereby achieving automatic classification of vector-borne disease samples. Specifically, the hardware environment requires an NVIDIA GPU supporting CUDA, and the software relies on tools such as Python (≥3.7) and PyTorch. The program receives training parameters via argparse. All input images undergo unified preprocessing before being fed into the model, including random cropping and numerical normalization, with a uniform input size of 224×224 to match the Swin-T network structure. The training and validation sets are loaded in mini-batch mode via DataLoader, with the training set shuffled and the validation set maintaining the original sample order. The model uses the Swin Transformer pre-trained on the ImageNet dataset provided by Torchvision as the backbone network, which possesses excellent general visual feature extraction capabilities. The initial Swin-T classification head is replaced with a fully connected layer adapted for the task, achieving mapping from high-dimensional visual features to target categories of multi-class vector-borne disease samples. Then, the cross-entropy loss function and the AdamW optimization algorithm are used. In each training batch, the operations of "forward propagation," "loss calculation," "backpropagation," "parameter update," and "learning rate update" are executed sequentially. Each epoch represents a complete traversal of the training set by the model. During training, the average training loss, Top-1 training accuracy, and current learning rate are calculated and recorded via TensorBoard. Finally, the validated classification results are generated and presented to the user in a report. The cross-entropy loss function is suitable for multi-class classification problems, directly measuring the difference between the predicted class probability distribution and the true label. The AdamW optimization algorithm, based on the adaptive learning rate, can also introduce weight decay, which helps improve the model's generalization ability.

[0066] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0067] The above embodiments are merely illustrative of the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be construed as limiting the scope of protection of the present invention. All equivalent changes or modifications made based on the essence of the content of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A device based on a vector-borne organism database, characterized in that, include: A fixed component includes a shelf on which a control bracket is provided; A data acquisition terminal is mounted on the control bracket, and the data acquisition terminal has an identification module, a marking module, a counting module, and a classification module; The process involves first placing the sample on the platform, then controlling the position of the acquisition terminal via the control bracket to sample the sample, and finally identifying the sample through the identification module, marking it through the tagging module, and classifying it through the classification module.

2. The device based on a vector-borne organism database according to claim 1, characterized in that, The control bracket includes a clamping bracket and a fixing bracket; The clamping bracket includes a telescopic robotic arm and a clamping component. The clamping component is disposed at one end of the telescopic robotic arm and consists of two parts: one part has a gear pin and the other part has a serrated groove. The gear pin is adapted to the serrated groove. The fixed bracket is equipped with a telescopic support rod. One end of the telescopic support rod is connected to the fixed bracket through a spherical structural member. The fixed bracket is also equipped with a threaded fastener. The surface of the telescopic support rod has a texture for matching the threaded fastener. The threaded fastener is also equipped with a screw knob. Rotating the screw knob controls the tightness between the telescopic support rod and the threaded fastener. The other end of the telescopic support rod is equipped with a plastic gasket.

3. The device based on a vector-borne organism database according to claim 1, characterized in that, The bottom of the display stand is provided with a sample tray, one end of which is provided with a pull ring, the two sides of the other end of the sample tray are inclined upwards, and an opening is provided in the middle of the other end of the sample tray.

4. The device based on a vector-borne organism database according to claim 1, characterized in that, The inner wall of the shelf is equipped with a surround-type variable temperature lamp, which includes an inner light strip and an outer diffuser. The outer diffuser is made of a semi-transparent frosted material.

5. The device based on a vector-borne organism database according to claim 1, characterized in that, The inner wall of the bracket connecting the shelf and the control bracket has a groove, and an embedded light is installed in the groove. The embedded light includes an inner light source and an outer light cover. The embedded light can be popped out or fixed by pressing, and the embedded light has a foldable and rotatable bracket.

6. The device based on a vector-borne organism database according to claim 2, characterized in that, The gear pin is cross-shaped, the thickness of the gear pin tip circle is 0.1mm, and the spacing between each tooth of the sawtooth groove is 0.15mm.

7. The device based on a vector-borne organism database according to claim 1, characterized in that, The data acquisition terminal is a mobile phone with a camera function, and the mobile phone has an application program installed that includes the identification module, the marking module, the counting module and the classification module.

8. A device application method based on a vector-borne organism database, characterized in that, Performed using the apparatus as described in any one of claims 1 to 7, including the following specific steps: Create a task queue; Obtain sample image information, identify the sample image information through a preset algorithm service layer, and put it into the task queue; Preprocess the sample image information in the task queue; Build a deep learning model, connect the deep learning model with a vector-borne organism database, and train it. The pre-processed sample image information is analyzed and compared using a trained deep learning model; The analysis and comparison results are used to generate a report, and key information is marked in the report.

9. The device application method based on a vector-borne organism database according to claim 8, characterized in that, After preprocessing the sample image information in the task queue, the following specific steps are also included: Construct a YoLo11n model; Load the weights from YoLon.pt to fine-tune the YoLo11n model; The fine-tuned YoLo11n model was tested in conjunction with the built deep learning detection framework; The tested YoLo11n model is imported into the pre-processed sample image information for counting; Generate a report from the counting results.

10. The device application method based on a vector-borne organism database according to claim 8, characterized in that, After preprocessing the sample image information in the task queue, the following specific steps are also included: Construct a multi-class image classification model; The multi-class image classification model was pre-trained on the ImageNet dataset. The pre-trained model is imported into the pre-processed sample image information for classification. The classification results were validated using the cross-entropy loss function and the AdamW optimization algorithm. Generate a report from the verified classification results.

Images