Method, system, apparatus and readable storage medium for gender differentiation of aedes albopictus

Abstract

The application discloses a method, system and device for gender distinguishing of Aedes albopictus and a readable storage medium, and comprises the following steps: calibrating and selecting regions of each picture in the obtained data set; calibrating key features of each Aedes albopictus picture in the Aedes albopictus detection data set; based on the gender features of Aedes albopictus, pre-processing the key features of each Aedes albopictus region picture in the Aedes albopictus region data set to form an Aedes albopictus gender classification data set; obtaining an Aedes albopictus detection model, a key feature position detection model and an Aedes albopictus gender classification model; and detecting a to-be-detected picture by using the Aedes albopictus detection model, the key feature position detection model and the Aedes albopictus gender classification model in sequence to obtain a detection result. The application reduces and saves time and economic cost. The relative quantity of Aedes albopictus and the change in the quantity of the gender types of the Aedes albopictus caught can be used for monitoring the outbreak of the Aedes albopictus in the region and thus timely adjusting corresponding prevention and control measures.

Description

Technical Field

[0001] This invention relates to the field of big data technology, and in particular to a method, system, device, and readable storage medium for sex differentiation of Aedes albopictus mosquitoes. Background Technology

[0002] Aedes albopictus, also known as the Asian tiger mosquito, belongs to the family Culicidae in the order Diptera and originates from Southeast Asia. In my country, it is commonly called the "flower mosquito" or "poisonous mosquito" due to its black and white striped markings. When it bites, it releases a toxin that causes redness, swelling, and itching around the bite site.

[0003] The impact of Aedes albopictus on humans is growing, exhibiting characteristics of an invasive species in my country. Its population is surging, expanding northward and making control extremely difficult. Moreover, Aedes albopictus not only feeds on human blood but also on the blood of other mammals and some birds. Even with human protection measures, it still survives in the wild. Furthermore, Aedes albopictus is highly adaptable, tolerating cold and becoming even more active under strong ultraviolet radiation.

[0004] With the increasing population of Aedes albopictus mosquitoes, sex determination has become crucial for their control. Traditional methods of sex determination almost always involve capturing mosquitoes and then visually identifying their sex. This method is extremely labor-intensive and highly susceptible to human error, which increases the difficulty of controlling Aedes albopictus and also delays the provision of results, thus increasing time costs. Summary of the Invention

[0005] This invention addresses the shortcomings of existing technologies by providing a method, system, device, and readable storage medium for sex differentiation of Aedes albopictus mosquitoes.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] A method for sexing Aedes albopictus mosquitoes includes the following steps:

[0008] The images in the obtained dataset were labeled and regions were selected to determine whether each image contained Aedes albopictus, thus forming an Aedes albopictus detection dataset.

[0009] Key features were identified for each Aedes albopictus image in the Aedes albopictus detection dataset to form an Aedes albopictus region dataset;

[0010] Based on the sex characteristics of Aedes albopictus, the key features of each Aedes albopictus region image in the Aedes albopictus region dataset are preprocessed to form an Aedes albopictus sex classification dataset.

[0011] Based on the Aedes albopictus detection dataset, Aedes albopictus region image dataset, and Aedes albopictus sex classification dataset, a pre-trained model for Aedes albopictus detection, a pre-trained model for key feature location detection, and a pre-trained model for Aedes albopictus sex classification were trained respectively, resulting in the Aedes albopictus detection model, the key feature location detection model, and the Aedes albopictus sex classification model.

[0012] The images to be detected were sequentially analyzed using the Aedes albopictus detection model, the key feature location detection model, and the Aedes albopictus sex classification model to obtain the detection results.

[0013] As one possible implementation method, the following steps are also included:

[0014] Obtain the dataset, where each image has a fixed size, specifically 5472x3648.

[0015] As one possible implementation, the step of labeling and selecting regions for each image in the obtained dataset to obtain an Aedes albopictus detection dataset includes the following steps:

[0016] Each image in the dataset is labeled to determine whether it contains Aedes albopictus, forming an Aedes albopictus detection dataset, which includes Aedes albopictus dataset and non-Aedes albopictus dataset.

[0017] Each image in the Aedes albopictus dataset is processed by taking a screenshot of the target region to obtain Aedes albopictus region images, thus forming an Aedes albopictus classification dataset. Each Aedes albopictus region image is the same size.

[0018] As one possible implementation, the step of identifying key features of each Aedes albopictus image in the Aedes albopictus detection dataset to form an Aedes albopictus region dataset includes the following steps:

[0019] Key features are identified in each Aedes albopictus region image to obtain feature region images, forming an Aedes albopictus region dataset. The key features include the antennae flagella, mouthparts, and tail of Aedes albopictus.

[0020] Furthermore, the specific location information of each region in the feature region image is saved to form a specific location information set, which includes the x-axis and y-axis values ​​of the upper left and lower right coordinates of the target box and the corresponding category name.

[0021] As one possible implementation, the preprocessing includes any one or two of the following steps:

[0022] The sharpness of key feature parts in each feature region image in the Aedes albopictus region dataset is corrected to obtain sharp key region images, forming a dataset to be classified. The correction process includes any one of gamma correction, brightness enhancement, and generating sharp data using adversarial networks.

[0023] Based on the dataset to be classified, after data balancing and selection, a sex classification dataset for Aedes albopictus mosquitoes was obtained.

[0024] As one possible implementation, the step of sequentially detecting the image to be detected using an Aedes albopictus detection model, a key feature location detection model, and an Aedes albopictus sex classification model to obtain the detection result includes the following steps:

[0025] The image to be detected is input into the Aedes albopictus detection model to determine whether the image contains Aedes albopictus.

[0026] If so, the result is input into the key feature location detection model to detect whether it contains key feature locations;

[0027] If so, the result is input into the Aedes albopictus sex classification model to obtain the Aedes albopictus sex classification.

[0028] As one possible implementation method, the Aedes albopictus detection model, the key feature location detection model, and the Aedes albopictus sex classification model are obtained through the following methods:

[0029] A pre-trained model for detecting Aedes albopictus was constructed, and the model was trained and tested based on the Aedes albopictus detection dataset to obtain the Aedes albopictus detection model.

[0030] A key feature location detection model is constructed. The key feature location detection model is trained and tested based on the key feature labeling data of the feature region detection dataset to obtain the key feature location detection model.

[0031] A pre-trained model for Aedes albopictus sex classification was constructed based on the efficientNet-b4 model. The pre-trained model was trained and tested using the Aedes albopictus sex classification dataset to obtain the Aedes albopictus sex classification model.

[0032] A sex differentiation system for Aedes albopictus mosquitoes includes a first processing module, a second processing module, a third processing module, a construction and training module, and a result detection module.

[0033] The first processing module is used to label and select regions for each image in the obtained dataset to determine whether it contains Aedes albopictus, thus forming an Aedes albopictus detection dataset;

[0034] The second processing module performs key feature identification on each Aedes albopictus image in the Aedes albopictus detection dataset to form an Aedes albopictus region dataset;

[0035] The third processing module preprocesses the key features of each Aedes albopictus region image in the Aedes albopictus region dataset based on the sex characteristics of Aedes albopictus, forming an Aedes albopictus sex classification dataset.

[0036] The construction training module trains a pre-trained model for Aedes albopictus detection, a pre-trained model for key feature location detection, and a pre-trained model for Aedes albopictus sex classification based on the Aedes albopictus detection dataset, the Aedes albopictus region image dataset, and the Aedes albopictus sex classification dataset, respectively, to obtain the Aedes albopictus detection model, the key feature location detection model, and the Aedes albopictus sex classification model.

[0037] The result detection module sequentially uses the Aedes albopictus detection model, the key feature location detection model, and the Aedes albopictus sex classification model to detect the image to be detected, and obtains the detection results.

[0038] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described above.

[0039] A sex differentiation device for Aedes albopictus mosquitoes includes a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the method described above.

[0040] This invention, by adopting the above technical solutions, has significant technical effects:

[0041] This invention reduces the workload of manual verification, saving time and economic costs. It continuously pushes information on Aedes albopictus mosquito identification in a region via the device, and monitors outbreaks by tracking changes in the relative numbers and sex-specific trapping quantities of Aedes albopictus, allowing for timely adjustments to control measures. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a schematic diagram of the overall process of the method of the present invention;

[0044] Figure 2This is a schematic diagram of the overall structure of the system of the present invention;

[0045] Figure 3 This is an example illustration showing the differences between the female and male insects of this invention;

[0046] Figure 4 This is a schematic diagram of the model composite scaling method in Efficient-Net;

[0047] Figure 5 and Figure 6 The following are schematic diagrams showing the recognition results. Detailed Implementation

[0048] The present invention will be further described in detail below with reference to the embodiments. The following embodiments are explanations of the present invention, but the present invention is not limited to the following embodiments.

[0049] A method for sexing Aedes albopictus mosquitoes, such as Figure 1 As shown, it includes the following steps:

[0050] S100. Label and select regions for each image in the obtained dataset to form an Aedes albopictus detection dataset;

[0051] S200. Key features are identified for each Aedes albopictus image in the Aedes albopictus detection dataset to form an Aedes albopictus region dataset.

[0052] S300. Based on the sex characteristics of Aedes albopictus, the key features of each Aedes albopictus region image in the Aedes albopictus region dataset are preprocessed to form an Aedes albopictus sex classification dataset.

[0053] S400. Based on the Aedes albopictus detection dataset, the Aedes albopictus region image dataset, and the Aedes albopictus sex classification dataset, a pre-trained model for Aedes albopictus detection, a pre-trained model for key feature location detection, and a pre-trained model for Aedes albopictus sex classification are trained respectively to obtain the Aedes albopictus detection model, the key feature location detection model, and the Aedes albopictus sex classification model.

[0054] S500 sequentially uses the Aedes albopictus detection model, the key feature location detection model, and the Aedes albopictus sex classification model to detect the image to be tested, and obtains the detection results.

[0055] In one embodiment, the method further includes the following step: obtaining a dataset in which each image has a fixed size, wherein the fixed size is 5472x3648.

[0056] In one embodiment, the step of labeling and selecting regions for each image in the obtained dataset to obtain an Aedes albopictus detection dataset includes the following steps:

[0057] Each image in the dataset is labeled to determine whether it contains Aedes albopictus, forming an Aedes albopictus detection dataset, which includes Aedes albopictus dataset and non-Aedes albopictus dataset.

[0058] Each image in the Aedes albopictus dataset is processed by taking a screenshot of the target region to obtain Aedes albopictus region images, thus forming an Aedes albopictus classification dataset. Each Aedes albopictus region image is the same size.

[0059] Specifically, the step of identifying key features for each Aedes albopictus image in the Aedes albopictus detection dataset to form an Aedes albopictus region dataset includes the following steps:

[0060] Key features are identified in each Aedes albopictus region image to obtain feature region images, forming an Aedes albopictus region dataset. The key features include the antennae flagella, mouthparts, and tail of Aedes albopictus.

[0061] Furthermore, the specific location information of each region in the feature region image is saved to form a specific location information set, which includes the x-axis and y-axis values ​​of the upper left and lower right coordinates of the target box and the corresponding category name.

[0062] In addition, the preprocessing includes any one or two of the following steps:

[0063] The sharpness of key feature parts in each feature region image in the Aedes albopictus region dataset is corrected to obtain sharp key region images, forming a dataset to be classified. The correction process includes any one of gamma correction, brightness enhancement, and generating sharp data using adversarial networks.

[0064] Based on the dataset to be classified, after data balancing and selection, a sex classification dataset for Aedes albopictus mosquitoes was obtained.

[0065] Finally, the detection results are obtained by sequentially using the Aedes albopictus mosquito detection model, the key feature location detection model, and the Aedes albopictus mosquito sex classification model to detect the images to be tested, including the following steps:

[0066] The image to be detected is input into the Aedes albopictus detection model to determine whether the image contains Aedes albopictus.

[0067] If so, the result is input into the key feature location detection model to detect whether it contains key feature locations;

[0068] If so, the result is input into the Aedes albopictus sex classification model to obtain the Aedes albopictus sex classification.

[0069] In one embodiment, the Aedes albopictus detection model, the key feature location detection model, and the Aedes albopictus sex classification model are obtained through the following methods:

[0070] A pre-trained model for detecting Aedes albopictus was constructed, and the model was trained and tested based on the Aedes albopictus detection dataset to obtain the Aedes albopictus detection model.

[0071] A key feature location detection model is constructed. The key feature location detection model is trained and tested based on the key feature labeling data of the feature region detection dataset to obtain the key feature location detection model.

[0072] A pre-trained model for Aedes albopictus sex classification was constructed based on the efficientNet-b4 model. The pre-trained model was trained and tested using the Aedes albopictus sex classification dataset to obtain the Aedes albopictus sex classification model.

[0073] The following example illustrates the processing of specific data:

[0074] The implementation environment of this application:

[0075] The instrument selected is the Top Cloud Agriculture Insect Intelligent Monitoring System. This device utilizes the phototaxis of mosquitoes, with a lamp wavelength of 350nm-450nm. It also incorporates mosquito attractants and is filled with a certain amount of carbon dioxide to simulate a human environment. The camera has 20 megapixels. A hardware structure diagram is attached. Figure 2 As shown.

[0076] Step 1: Identify the insect type and location;

[0077] 1. Image capture by the device: Using a 20-megapixel lens, live mosquitoes were photographed, with 2 images taken per minute, for a total of 10,000 training images, representing approximately 20,000 targets. Each image was 5472x3648 pixels in size.

[0078] 2. Creating an Aedes albopictus detection dataset: Real images of Aedes albopictus were collected, and each image was labeled using a calibration tool to create an Aedes albopictus detection dataset. This dataset includes two categories: Aedes albopictus and non-Aedes albopictus. The dataset contains approximately 12,000 Aedes albopictus targets, with the remaining 8,000 targets representing other mosquitoes or non-mosquito categories. Of these, approximately 10,000 are female Aedes albopictus and 2,000 are males.

[0079] 3. Generate a classification dataset using the calibration file: Based on the Aedes albopictus detection dataset, the calibration file is used to crop the target area of ​​each Aedes albopictus using the cv2.crop function, forming 12,000 close-up images of Aedes albopictus. Each small image is about 400x400 pixels in size. These small images are collected to form the original classification dataset.

[0080] 4. Training the Detection Model: The data is based on the Aedes albopictus mosquito detection dataset. The YOLOv7 training framework is used to train the Aedes albopictus detection model. The YOLOv7 detection framework used in this embodiment integrates the advantages of YOLOv5, YOLO-R, and YOLO-X network frameworks and adds some new techniques to achieve current state-of-the-art (SOTA) performance. It inherits the overall network architecture and configuration files of YOLOv5 while maintaining the same training, inference, and testing processes. It also borrows network design methods, hyperparameter setting methods, and implicit knowledge learning methods from YOLO-R, while adopting YOLOv-X's dynamic label allocation method (SimOTA). After integrating the advantages of the three frameworks, YOLOv7 proposes efficient aggregation networks ELAN and E-ELAN, reparameterized convolutional RepConv, auxiliary head detection, and model scaling to balance speed, accuracy, and complexity. YOLOv7 can maintain ultra-high FPS while maintaining optimal accuracy, making it the fastest and most accurate detector currently available.

[0081] In this embodiment, the image input size is adjusted to 1600x1600, the output recognition model size is 75.1MB, the recognition speed is 161fps, and the recognition time per image is 2.8 milliseconds. The mAP on the validation set is 97.8%.

[0082] Step 2: Data Processing in Key Areas

[0083] 1. Identify key areas: After classifying the original data dataset, select male and female Aedes albopictus mosquitoes. Select 10,000 female mosquitoes and 2,000 male mosquitoes to observe the differences between the male and female mosquitoes.

[0084] This invention reveals that data captured by the device clearly shows significant differences in the morphology of the antennae flagella, mouthparts, and genitalia of male and female Aedes albopictus. Through literature review and expert consultation, it was found that males and females of Aedes albopictus can be distinguished by these three differences. Therefore, the antennae flagella, mouthparts, and genitalia were identified as three key areas for further calibration.

[0085] Antennae flagellum morphology difference: Male Aedes albopictus mosquitoes have dense, brush-like antennae flagels. Female mosquitoes have only short, sparse antennae flagels.

[0086] Differences in mouthpart morphology: Male Aedes albopictus mosquitoes have trident-shaped mouthparts. Females have needle-shaped mouthparts suitable for piercing and sucking blood.

[0087] Tail structure difference: Male Aedes albopictus mosquitoes have a tail-like reproductive organ. The female mosquito in the image does not have this structure on its tail; the female's reproductive organ is generally hidden.

[0088] Examples of differences between female and male insects can be seen in the attached image. Figure 3 .

[0089] 2. Forming a Feature Region Detection Dataset: Based on the above distinctions, and using small images of Aedes albopictus mosquitoes, these three feature points are labeled to form a feature region detection dataset. The detailed labeling process involves using the labelimg software tool to label three regions on the small images of Aedes albopictus mosquitoes: mouthparts, tail, and antennal flagella. The resulting XML file includes the specific location information for each category, including the x-axis and y-axis values ​​of the top-left and bottom-right coordinates of the target bounding box, and the corresponding category name. Due to the imbalance in the number of male and female mosquitoes, this invention balances the number of male and female Aedes albopictus mosquitoes and selects 1500 females and 1500 males as the base images for labeling, resulting in the labeling of 4500 mouthpart targets, 4500 antennal flagella targets, and 4500 tail genital structure targets.

[0090] 3. Training the Detection Model: A key feature location detection model was trained using the feature region detection dataset. Based on the feature region detection dataset, the YOLOv7 training framework was used to train the key feature location detection model. Due to the relatively small size of the feature region detection dataset, the input size of the key feature location detection model was set to 416x416, resulting in an output size of 75.1MB. The recognition speed was 261fps, with a single image recognition time of 1.8 milliseconds, and an mAP of 99.1% on the validation set.

[0091] 4. Improve the clarity of key regions in the feature region detection dataset to obtain the dataset to be classified: Due to the delay of the shooting camera and the high vibration frequency of mosquitoes, some images may not be clear enough, resulting in the three features not being obvious. Therefore, this embodiment uses gamma correction, brightness enhancement and adversarial network to generate clear data to improve the image clarity, improve the brightness and clarity of key region images, and achieve the purpose of improving the local features of key regions.

[0092] The significance of improving the clarity of key regions in feature region detection datasets lies in:

[0093] 1. Instead of processing the entire image, only three key areas are processed, which can improve the image processing speed compared to processing the entire image.

[0094] 2. After image sharpness processing in certain areas, these features can be used as classification features input into the classification network;

[0095] After performing local area detection, the overall image brightness is too dark. The purpose of image gamma value correction and brightness enhancement is to improve brightness and highlight feature areas.

[0096] Furthermore, a dataset of 3000 sets was created by pairing a low-quality image with a high-quality image as training data. A GAN (Generative Adversarial Network) model was trained using this approach. This model's function is to take a low-quality image as input and output a high-quality image. Inputting some low-quality images from classification dataset 1 into this model yielded a batch of relatively high-quality data. These images, after image processing transformation and the addition of the adversarial network, constitute the foundation dataset for the dataset to be classified.

[0097] Step 3: Gender Recognition

[0098] 1. Create the dataset to be classified:

[0099] The data was generated by improving the image clarity of the base dataset to be classified. After data balancing and selection, 1500 male insect data and 1500 female insect data were selected and collected to obtain the dataset to be classified.

[0100] 2. Training the sex classification model: On the dataset to be classified, the Aedes albopictus sex classification model is trained using the efficientNet-b4 model.

[0101] EfficientNet effectively balances the three dimensions of depth, width, and resolution by uniformly scaling these three dimensions using a fixed set of scaling factors. EfficientNet designs a standardized convolutional network expansion method that not only achieves high accuracy but also significantly saves computational resources. EfficientNet proposes the following... Figure 4 The model composite scaling method.

[0102] The classification accuracy of the dataset to be classified on the efficientNet-b4 model can reach 99%.

[0103] 3. Build the program and test its effectiveness.

[0104] The test used 500 real, untrained images, employed an algorithm program for testing, and was manually verified to obtain the final recognition results.

[0105] The detailed recognition results are shown in the table below:

[0106] Algorithm recognition Manual identification accuracy Aedes albopictus 698 723 96.5％ male Aedes albopictus mosquitoes 78 88 88.6％ Female Aedes albopictus mosquito 620 635 97.6％

[0107] The Aedes albopictus mosquito body localization and recognition rate reached 96.5% in the test set. The algorithm's accuracy in identifying male mosquitoes reached 88.6%, with errors mainly due to unclear features caused by mosquito vibration or occlusion of features. The recognition rate for female Aedes albopictus mosquitoes reached 97.6%, with errors mainly due to feature occlusion caused by mosquito accumulation.

[0108] This application includes a detection model for Aedes albopictus, a key feature location detection model, and a sex classification model;

[0109] Aedes albopictus detection model: used to locate Aedes albopictus mosquitoes in the original image and extract the mosquito area.

[0110] Key Feature Location Detection Model: This model identifies three key regions within the insect body area generated by Detection Model 1: mouthparts, flagella, and tail. Image sharpness in these three key regions is adjusted using image processing and a GAN (Google Approach). The generated images serve as the basis for the classification model.

[0111] Sex classification model: used to distinguish the sex of Aedes albopictus mosquitoes.

[0112] Example 2:

[0113] A sex differentiation system for Aedes albopictus mosquitoes, such as Figure 2 As shown, it includes a first processing module 100, a second processing module 200, a third processing module 300, a construction and training module 400, and a result detection module 500;

[0114] The first processing module 100 is used to label and select regions for each image in the obtained dataset to determine whether it contains Aedes albopictus, thereby forming an Aedes albopictus detection dataset;

[0115] The second processing module 200 performs key feature labeling on each Aedes albopictus image in the Aedes albopictus detection dataset to form an Aedes albopictus region dataset;

[0116] The third processing module 300 preprocesses the key features of each Aedes albopictus region image in the Aedes albopictus region dataset based on the sex characteristics of Aedes albopictus, forming an Aedes albopictus sex classification dataset.

[0117] The construction training module 400 trains a pre-training model for Aedes albopictus detection, a pre-training model for key feature location detection, and a pre-training model for Aedes albopictus sex classification based on the Aedes albopictus detection dataset, the Aedes albopictus region image dataset, and the Aedes albopictus sex classification dataset, respectively, to obtain the Aedes albopictus detection model, the key feature location detection model, and the Aedes albopictus sex classification model.

[0118] The result detection module 500 sequentially detects the image to be detected using the Aedes albopictus detection model, the key feature location detection model, and the Aedes albopictus sex classification model to obtain the detection results.

[0119] As the device embodiment is basically similar to the method embodiment, the description is relatively simple, and relevant parts can be found in the description of the method embodiment.

[0120] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0121] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0122] This invention is described with reference to flowchart illustrations and / or block diagrams of the method, terminal device (system), and computer program product according to the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing terminal device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing terminal device, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0123] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing terminal device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0124] These computer program instructions can also be loaded onto a computer or other programmable data processing terminal equipment, causing a series of operational steps to be performed on the computer or other programmable terminal equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable terminal equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0125] Furthermore, it should be noted that the shapes and names of the parts and components described in the specific embodiments described in this specification may differ. All equivalent or simple variations made to the structure, features, and principles described in this patent concept are included within the protection scope of this patent. Those skilled in the art to which this invention pertains may make various modifications or additions to the described specific embodiments or use similar methods to replace them, as long as they do not depart from the structure of this invention or exceed the scope defined in these claims, they should all fall within the protection scope of this invention.

Claims

1. A method of gender differentiation of Aedes albopictus, characterized in that, The method comprises the following steps: The obtained data set is labeled and regionally selected to determine whether each picture contains Aedes albopictus, thereby forming an Aedes albopictus detection data set; Each Aedes albopictus picture in the Aedes albopictus detection data set is labeled for key features, thereby forming an Aedes albopictus region data set, which comprises the following steps: The key features of each Aedes albopictus region picture are labeled to obtain a feature region image, thereby forming an Aedes albopictus region data set, wherein the key features include the antennal flagellum, mouthparts and tail of the Aedes albopictus; The specific position information of each region in the feature region image is saved to form a specific position information set, and the specific position information includes the x-axis and y-axis numerical values of the upper left coordinates of the target frame, the x-axis and y-axis numerical values of the lower right coordinates and the corresponding category name; Based on the gender characteristics of the Aedes albopictus, the key features of each Aedes albopictus region picture in the Aedes albopictus region data set are preprocessed to form an Aedes albopictus gender classification data set, and the preprocessing comprises any one or two of the following steps: The clarity of the key feature parts in each feature region image in the Aedes albopictus region data set is corrected to obtain a clear key region image, thereby forming a classification data set, wherein the correction processing includes any one of gamma correction, brightness enhancement and adversarial network generated clear data; Based on the classification data set, the Aedes albopictus gender classification data set is obtained after data balancing selection; Based on the Aedes albopictus detection data set, the Aedes albopictus region picture data set and the Aedes albopictus gender classification data set, an Aedes albopictus detection pre-training model, a key feature position detection pre-training model and an Aedes albopictus gender classification pre-training model are trained, thereby obtaining an Aedes albopictus detection model, a key feature position detection model and an Aedes albopictus gender classification model; The detection picture is sequentially detected by the Aedes albopictus detection model, the key feature position detection model and the Aedes albopictus gender classification model, thereby obtaining a detection result.

2. The method for sex differentiation of Aedes albopictus mosquitoes according to claim 1, characterized in that, The method further comprises the following steps: A data set is obtained, and the size of each picture in the data set is a fixed size, wherein the fixed size is 5472x3648.

3. The method for sex differentiation of Aedes albopictus mosquitoes according to claim 1, characterized in that, The obtained data set is labeled and regionally selected to determine whether each picture contains Aedes albopictus, thereby forming an Aedes albopictus detection data set, which comprises the following steps: Each picture in the data set is labeled to determine whether it contains Aedes albopictus, thereby forming an Aedes albopictus detection data set, wherein the Aedes albopictus detection data set includes an Aedes albopictus data set and a non-Aedes albopictus data set; Each picture in the Aedes albopictus data set is subjected to target region screenshot processing to obtain Aedes albopictus region pictures to form an Aedes albopictus classification data set, and the size of each Aedes albopictus region picture is the same.

4. The method for sex differentiation of Aedes albopictus mosquitoes according to claim 1, characterized in that, The detection picture is sequentially detected by the Aedes albopictus detection model, the key feature position detection model and the Aedes albopictus gender classification model, thereby obtaining a detection result, which comprises the following steps: The detection picture is input into the Aedes albopictus detection model to determine whether the detection picture contains Aedes albopictus; If yes, the result is input into the key feature position detection model to determine whether the key feature position is contained; If yes, the result is input into the Aedes albopictus gender classification model to obtain the Aedes albopictus gender classification.

5. The method for sex differentiation of Aedes albopictus mosquitoes according to claim 1, characterized in that, The Aedes albopictus detection model, the key feature position detection model and the Aedes albopictus gender classification model are obtained by the following manner: An Aedes albopictus detection pre-training model is constructed, and the Aedes albopictus detection pre-training model is trained and tested based on an Aedes albopictus detection dataset to obtain the Aedes albopictus detection model; An Aedes albopictus detection pre-training model is constructed, and the Aedes albopictus detection pre-training model is trained and tested based on an Aedes albopictus detection dataset to obtain the Aedes albopictus detection model; An Aedes albopictus gender classification pre-training model is constructed based on an efficientNet-b4 model, and the Aedes albopictus gender classification pre-training model is trained and tested based on an Aedes albopictus gender classification dataset to obtain the Aedes albopictus gender classification model.

6. A system for gender differentiation of Aedes albopictus, comprising: The first processing module, the second processing module, the third processing module, the construction training module and the result detection module are included. The first processing module is configured to calibrate and select regions of whether each picture in the obtained dataset contains Aedes albopictus to form an Aedes albopictus detection dataset. The second processing module is configured to calibrate key features of each Aedes albopictus picture in the Aedes albopictus detection dataset to form an Aedes albopictus region dataset, including the following steps: In each Aedes albopictus region picture, key features are calibrated to obtain a feature region image to form the Aedes albopictus region dataset, wherein the key features include an antenna flag, a mouth organ and a tail of the Aedes albopictus; The specific position information of each region in the feature region image is saved to form a specific position information set, and the specific position information includes an x-axis and a y-axis value of a top-left coordinate of a target frame, an x-axis and a y-axis value of a bottom-right coordinate and a corresponding category name; The third processing module is configured to preprocess the key features of each Aedes albopictus region picture in the Aedes albopictus region dataset based on gender characteristics of the Aedes albopictus to form an Aedes albopictus gender classification dataset, and the preprocessing includes any one or two of the following steps: The clarity of the key feature part in each feature region image in the Aedes albopictus region dataset is corrected to obtain a clear key region image to form a classification dataset, and the correction processing includes any one of gamma correction, brightness enhancement and adversarial network generated clear data; After data balancing selection based on the classification dataset, the Aedes albopictus gender classification dataset is obtained; The construction training module is configured to train an Aedes albopictus detection pre-training model, a key feature position detection pre-training model and an Aedes albopictus gender classification pre-training model based on the Aedes albopictus detection dataset, the Aedes albopictus region picture dataset and the Aedes albopictus gender classification dataset respectively to obtain the Aedes albopictus detection model, the key feature position detection model and the Aedes albopictus gender classification model; The result detection module is configured to detect a to-be-detected picture by the Aedes albopictus detection model, the key feature position detection model and the Aedes albopictus gender classification model in sequence to obtain a detection result.

7. A computer-readable storage medium storing a computer program, wherein the computer program comprises the following steps of: receiving a request for a resource from a client; determining whether the client is authorized to access the resource; and if the client is authorized to access the resource, providing the resource to the client. The computer program, which is executed by a processor, implements the method as claimed in any one of claims 1 to 5.

8. A white-lined Aedes aegypti gender differentiation device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein, The processor, when executing the computer program, implements the method as claimed in any one of claims 1 to 5.

Images