Image-driven tick detection comb fused with deep learning model

Abstract

The invention discloses an image-driven tick detection comb fused with a deep learning model, which relates to the technical field of tick detection and comprises an initial exposure time configuration module, an image data acquisition module, a feature extraction and fuzzy quantization module, an ambiguity prediction module, an image classification module, a normal image processing module and a fuzzy image exposure adjustment module. The initial exposure time configuration module is used for firstly determining initial exposure time adaptive to the current environment and the detection requirement, and carrying out parameter configuration on the micro camera; according to the method, the image blurring degree is accurately predicted through deep learning and image processing technologies, normal and blurred images are distinguished, adaptive exposure time regulation is adopted for the blurred images, motion blurring is reduced, image definition is ensured, feature extraction accuracy is improved, the misjudgment probability is reduced, ticks are detected and removed in time, and the risk that a host infects pathogens is reduced. The robustness and reliability of the equipment in different environments are enhanced through dynamic exposure adjustment, and the detection precision is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of tick detection, and particularly to an image-driven tick detection comb integrated with a deep learning model. Background Art

[0002] An image-driven tick detection comb is an innovative device integrating computer vision and intelligent detection technologies, dedicated to efficiently identifying and removing ticks from the fur of pets or livestock. This comb embeds a micro camera or image sensor on the basis of traditional teeth, and can capture the images on the fur surface in real time during the combing process, and analyze the image features through the built-in machine learning algorithm to accurately identify the presence of ticks. Once a tick is detected, the system can transmit it to a mobile application through an LED indicator light, vibration reminder or wireless transmission, so that the user can quickly take cleaning measures. In addition, some high-end models are also equipped with automatic collection or electric shock removal functions to improve the efficiency of removing ticks and reduce the risk of human-pet infection. This device is applicable to household pet care, livestock epidemic prevention and parasite monitoring in the wild, significantly improving the convenience and accuracy of tick detection.

[0003] Image-driven tick detection generally includes several key steps. First is data acquisition, where images of the fur of pets or livestock are taken in real time through a micro camera (or image sensor). Next, image preprocessing is performed, including denoising, contrast enhancement, etc., to improve the quality and clarity of the images. Then, key features in the images are extracted, such as the shape, size, texture and color features of ticks, and these features are usually extracted through image segmentation techniques, edge detection algorithms or deep learning models. After feature extraction, machine learning algorithms (such as convolutional neural network CNN) are used to classify the extracted features to determine whether there are ticks in the image. If a tick is detected, the system will generate feedback information, such as highlighting the position of the tick, and prompting the user through vibration, sound or a display screen.

[0004] The prior art faces a challenge when extracting tick features: When ticks are externally stimulated (such as comb teeth contact or hair vibration), they will quickly move to avoid detection. Currently, the micro cameras of detection systems usually use a constant exposure time setting, which may result in a blurred image of the moving tick in the image when the exposure time is long, thus affecting the accuracy of feature extraction. As a result, ticks may be misjudged as hair nodules or dirt, leading to detection failure. This situation not only delays the timely removal of ticks, but also may increase the risk of the host being infected with pathogens.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure, and thus it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The object of the present invention is to provide an image-driven tick detection comb integrated with a deep learning model. By combining the deep learning model and image processing technology, it can accurately predict the image blurriness, effectively distinguish between normal images and blurred images. For blurred images, an adaptive exposure time control technology is adopted to reduce the motion blur effect caused by long exposure, thereby ensuring image clarity and improving the accuracy of feature extraction, reducing the probability of misjudgment, ensuring that ticks can be detected and removed in time, and effectively reducing the risk of host pathogen infection. At the same time, the dynamic exposure time adjustment enables the image-driven tick detection comb to operate stably under various environmental conditions, enhancing the robustness and reliability of the device, solving the problem of image blurring caused by the rapid movement of ticks, and greatly improving the overall accuracy of the detection system, so as to solve the problems in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions: An image-driven tick detection comb integrated with a deep learning model, including an initial exposure time configuration module, an image data acquisition module, a feature extraction and blurriness quantification module, a blurriness prediction module, an image classification module, a normal image processing module, and a blurred image exposure adjustment module; The initial exposure time configuration module first determines the initial exposure time suitable for the current environment and detection requirements, and configures the parameters of the micro camera; The image data acquisition module, after taking a photo, comprehensively obtains various information of the image, providing complete data support for subsequent feature extraction; The feature extraction and blurriness quantification module, after obtaining the complete image data, extracts the feature information related to image blurring through image processing technology, performs feature engineering processing on the extracted feature information, and preliminarily quantifies the image distortion caused by motion blur, providing a basis for subsequent blurriness judgment; The blurriness prediction module inputs the blurred features after feature engineering processing into a pre-trained deep learning model, and predicts the blurriness of the current image through the deep learning model; The image classification module divides the captured image into a normal image and a blurred image according to the result predicted by the deep learning model; The normal image processing module, for normal images, continues to keep the initial exposure time to expose the micro camera when taking pictures, and performs subsequent tick feature extraction, identification, and positioning; The blurred image exposure adjustment module, for blurred image, according to the predicted result of the image blur degree, uses the fuzzy logic algorithm to adaptively regulate the initial exposure time, shortens the actual exposure duration when the micro camera takes pictures, reduces the blur effect caused by movement, and then performs subsequent tick feature extraction, recognition and positioning when the image is determined to be a normal image.

[0008] Preferably, the initial exposure time suitable for the current environment and detection requirements is determined based on the static detection requirements. The static detection requirements mean that when detecting ticks, it is preset that the ticks and their surrounding environment are in a relatively static state to ensure that the micro camera can capture clear and motion-blur-free images.

[0009] Preferably, feature information related to image blur is extracted through image processing technology. Among them, the extracted feature information includes the gradient change of the image edge and the distribution of the image spatial texture. Feature engineering processing is performed on the extracted feature information to generate an edge gradient reference value and a spatial texture disorder reference value respectively, and the image distortion caused by motion blur is preliminarily quantified through the edge gradient reference value and the spatial texture disorder reference value, providing a basis for subsequent blur degree judgment.

[0010] Preferably, the edge gradient reference value and the spatial texture disorder reference value after feature engineering processing are input into a pre-trained deep learning model, and a blur quantization coefficient is generated based on the deep learning model, and the blur degree of the current image is predicted through the blur quantization coefficient.

[0011] Preferably, the blur quantization coefficient generated when predicting the blur degree of the current image through a pre-trained deep learning model is compared and analyzed with a preset blur quantization coefficient reference threshold, and the captured image is divided. The specific division steps are as follows: If the generated blur quantization coefficient is greater than the blur quantization coefficient reference threshold, the captured image is divided into a blurred image; If the generated blur quantization coefficient is less than or equal to the blur quantization coefficient reference threshold, the captured image is divided into a normal image.

[0012] Preferably, the specific steps for generating the edge gradient reference value by performing feature engineering processing on the gradient change feature of the extracted image edge are as follows: The second derivative of the Laplace operator is used to extract the gradient information of each point in the image, the gradient calculation of the image is performed, and the local brightness change of each pixel point in the image is captured. The extracted expression is as follows: , where: represents the image at the position The Laplacian operator result at a certain position represents the change in the brightness change rate of the pixel point, which can describe the bending degree of the image in the horizontal and vertical directions, and then capture the edge information of the image. Represents the pixel value of the image at the position ; and represent the second-order derivatives in the horizontal and vertical directions respectively; Based on the gradient change of each pixel in the image, an edge gradient reference value is defined to quantify the blurriness of the image. The edge gradient reference value reflects the overall blurriness of the image by accumulating the intensity of the gradient change and through an exponential decay function. The specific calculation expression is: , where represents the edge gradient reference value, represents the set of all pixel points in the image, is the absolute value of the gradient.

[0013] Preferably, the specific steps for generating a spatial texture disorder reference value by performing feature engineering on the distribution characteristics of the extracted image spatial texture are as follows: Use wavelet transform to decompose the image into multiple frequency sub-bands, and obtain local texture features at different levels through the decomposed frequency sub-bands. The texture information of each local area is represented as a feature vector, reflecting the local texture pattern of the area. To further quantify the disorder of these local texture features, a new local feature calculation formula is adopted, and the calculation expression is: , where: represents the local texture change degree at the position in the image, reflecting the difference in texture between the surrounding area and the central area of this position, represents the texture value at the position in the image, represents the texture value at the center of the local area, is a parameter for adjusting the texture change sensitivity; After completing the extraction of local texture features, the disorder degree of the image is quantified by calculating the local texture change degree of each local area, and a spatial texture disorder reference value is generated based on the quantification result of the local texture change degree to evaluate the overall texture disorder of the image. The generation expression of the spatial texture disorder reference value is: , where: is the quantified spatial texture disorder reference value, is the local texture feature at the position in the image, is the global texture average feature of the image, is a parameter for controlling the perception of texture differences, k is the index of the current processed local area, and as the calculation process progresses,k It will traverse all local regions, N is the total number of local regions divided in the image.

[0014] Preferably, for blurred image, according to the prediction result of the image blur degree, an adaptive regulation of the initial exposure time is performed using a fuzzy logic algorithm. The specific steps are as follows: When processing a blurred image, first, the blur quantization coefficient of the image is calculated through a deep learning model , to quantify the blur degree of the image. In order to perform effective exposure adjustment on the image, the blur quantization coefficient is compared with a preset reference threshold of the blur quantization coefficient , and based on the fuzzy logic algorithm, an exposure adjustment factor is calculated according to the difference between the two . The specific calculation formula for the exposure adjustment factor is: , where is the gain coefficient of fuzzy inference, which is used to adjust the sensitivity of exposure time adjustment; Based on the generated exposure adjustment factor , the initial exposure time is adjusted to optimize the shooting quality. The specific adjustment expression is: , where is the adjusted exposure time, that is, the actually used exposure time, is the initial exposure time, is the exposure adjustment coefficient, which controls the amplitude of the exposure time change.

[0015] In the above technical solution, the technical effects and advantages provided by the present invention are as follows: By combining a deep learning model and image processing technology, the present invention accurately predicts the image blur degree, effectively distinguishes between normal images and blurred images. For blurred images, an adaptive exposure time regulation technology is adopted, reducing the motion blur effect caused by long-time exposure, thereby ensuring image clarity, improving the accuracy of feature extraction, reducing the probability of misjudgment, ensuring that ticks can be detected and removed in time, and effectively reducing the risk of pathogen infection of the host. At the same time, the dynamic exposure time adjustment enables the image-driven tick detection comb to operate stably under various environmental conditions, enhancing the robustness and reliability of the device, solving the problem of image blur caused by the rapid movement of ticks, and greatly improving the overall accuracy of the detection system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings.

[0017] Figure 1 This is a schematic diagram of the modules of the image-driven tick detection comb integrating a deep learning model according to the present invention.

[0018] Figure 2 This is a schematic diagram of the principle of the image-driven tick detection comb according to the present invention. Specific embodiments

[0019] Example embodiments will now be described more fully with reference to the accompanying drawings. However, the example embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that this disclosure will be more complete and comprehensive, and will fully convey the concept of the example embodiments to those skilled in the art.

[0020] The present invention provides an image-driven tick detection comb integrating a deep learning model as shown in Figure 1 , which includes an initial exposure time configuration module, an image data acquisition module, a feature extraction and fuzzy quantization module, a blurriness prediction module, an image classification module, a normal image processing module, and a blurred image exposure adjustment module; The initial exposure time configuration module first determines the initial exposure time suitable for the current environment and detection requirements, and configures the parameters of the micro camera; This step ensures that the micro camera can capture images with appropriate exposure parameters during shooting, thereby balancing the acquisition of light and details and providing a clear image basis. The correct initial exposure setting is a prerequisite for subsequent image processing and detection, and can avoid problems such as over-dark or over-bright caused by improper exposure to a certain extent.

[0021] The initial exposure time suitable for the current environment and detection requirements is determined based on static detection requirements. The static detection requirements refer to the fact that when detecting ticks, it is preset that the ticks and their surrounding environment are in a relatively static state to ensure that the micro camera can capture clear and motion-blur-free images. This is convenient for accurately detecting and identifying non-moving ticks.

[0022] Before image acquisition, a suitable exposure time is reasonably set according to the lighting conditions of the detection environment, the characteristics of the detection object, and the hardware parameters of the micro camera to ensure the best balance of image quality. The exposure time refers to the time when the photosensitive element of the micro camera is exposed to light, which directly affects the brightness, clarity, and dynamic range of the image. If the exposure time is too short, the image may be too dark, resulting in missing details; if the exposure time is too long, it may cause overexposure or blurring or ghosting when the detection object moves slightly. Therefore, on the premise of static detection, an initial exposure time that can ensure sufficient brightness while minimizing noise and blurring as much as possible is required, so that subsequent feature extraction and tick recognition can be carried out based on high-quality images, improving the accuracy and stability of detection.

[0023] The image data acquisition module, after taking a photo, comprehensively obtains various information of the image, including raw pixel data, brightness distribution, color channels, image metadata, and other auxiliary information. The role of this step is to provide complete data support for subsequent feature extraction, enabling subsequent algorithms to make full use of every part of the image information to determine the degree of blurring.

[0024] The feature extraction and blur quantification module, after obtaining the complete image data, extracts feature information related to image blur through image processing techniques (such as edge detection, texture analysis, and frequency domain conversion), performs feature engineering processing on the extracted feature information, and initially quantifies the image distortion caused by motion blur, providing a basis for subsequent blur degree judgment; Extract feature information related to image blur through image processing techniques. Among them, the extracted feature information includes the gradient change of the image edge and the distribution of the image spatial texture. Perform feature engineering processing on the extracted feature information to generate an edge gradient reference value and a spatial texture disorder reference value respectively. Initially quantify the image distortion caused by motion blur through the edge gradient reference value and the spatial texture disorder reference value, providing a basis for subsequent blur degree judgment.

[0025] When performing tick detection based on image drive, a smaller gradient change in the image edge usually means a higher degree of image blur. When there is motion blur or other types of blur in the image, the originally clearly visible edges and details become blurred, resulting in a smoother pixel gradient (i.e., brightness change) in the image. This is because the blurring process makes the transition of the edge less abrupt, thus reducing the sharpness of the edge and the intensity of the gradient. In a normal clear image, the gradient change of the edge is large, and the object and the background can be clearly distinguished. By calculating the gradient change in each area of the image, the blur degree of the image can be quantified. An image with a smaller edge gradient change usually indicates a higher degree of blur.

[0026] The specific steps for performing feature engineering on the gradient change features of the extracted image edges to generate an edge gradient reference value are as follows: First, perform gradient calculation on the image. The purpose is to capture the local brightness changes of each pixel in the image, especially in the edge regions. Use high-order partial derivatives or the Laplacian operator to extract the gradient information of each point in the image. Here, the second-order derivative of the Laplacian operator is used to extract the gradient information of each point in the image, and the extraction expression is as follows: , where: represents the image at position The result of the Laplacian operator, specifically the second-order derivative at that position. That is to say, it represents the change in the brightness change rate of that pixel point, can describe the bending degree of the image in the horizontal and vertical directions, and thus capture the edge information of the image, represents the pixel value of the image at position , and represent the second-order derivatives in the horizontal and vertical directions respectively; The function of this step is to calculate the brightness change of each point in the image. Especially at the edges, the brightness change is large, which helps to detect the edge information of the image.

[0027] Based on the gradient changes of each pixel in the image, define an edge gradient reference value to quantify the blurriness of the image, especially the distortion caused by motion blur. The edge gradient reference value can be obtained by accumulating the intensity of the gradient changes and using an exponential decay function to reflect the overall blurriness of the image. The specific calculation expression is: , where, represents the edge gradient reference value, represents the set of all pixel points in the image, is the absolute value of the gradient. This formula combines the intensity of the gradient changes with its exponential decay, and can more sensitively capture the weak gradient change regions caused by motion blur, thereby quantifying the blurriness of the image. The larger the edge gradient reference value, the clearer the edges in the image and the lower the blurriness; conversely, the smaller the edge gradient reference value, the higher the blurriness of the image.

[0028] From the edge gradient reference value, it can be seen that the larger the performance value of the edge gradient reference value generated after feature engineering processing of the gradient change characteristics of the extracted image edge, the higher the clarity of the image; conversely, the higher the blurriness of the image. The edge gradient reference value quantifies the edge sharpness of the image by calculating the gradient changes of each pixel point in the image and combining exponential decay. In a clear image, the edge is usually sharp, the pixel value changes are relatively drastic, the gradient value is large, resulting in a higher edge gradient reference value. In a blurred image, the edge becomes blurred, the pixel value changes are gentle, the gradient is small, thus making the edge gradient reference value lower. Therefore, the edge gradient reference value can effectively reflect the blurriness of the image. The smaller the value, the blurrier the image; the larger the value, the clearer the image.

[0029] When performing tick detection based on image driving, a higher degree of disorder in the distribution of the image spatial texture usually indicates a higher degree of blurriness of the image. In image processing, the texture distribution usually reflects the richness of details in the image. When the image is clear, the details and textures have a relatively regular and obvious distribution, the edges are clear, and the changes within the local area are relatively orderly. On the contrary, when the image is blurred, the details are smoothed or blurred, the texture distribution in the image becomes disordered and chaotic, the local changes are relatively smooth and lack obvious edges and features. Especially in the case of motion blur, the directionality and structure of the texture may disappear, resulting in a significant increase in the disorder of the texture in the image. Therefore, the disorder of the spatial texture can be used as an effective indicator of image blurriness, indicating a higher degree of blurriness in the image and obvious loss of details.

[0030] The specific steps for generating the spatial texture chaos reference value by performing feature engineering processing on the distribution characteristics of the extracted image spatial texture are as follows: First, perform multi-scale wavelet transform or local binary pattern (LBP) extraction on the image to obtain the texture features within each local area. Through these methods, the tiny texture changes in the image can be captured, especially in the high-frequency and detailed parts. Use wavelet transform (such as Haar wavelet or Daubechies wavelet) to decompose the image into multiple frequency sub-bands, and obtain the local texture features at different levels through these frequency sub-bands. The texture information of each local area can be represented as a feature vector, reflecting the local texture pattern of that area. To further quantify the disorder of these local texture features, a new local feature calculation formula is adopted, and the calculation expression is: , where: represents the local texture change degree at position in the image, reflecting the difference in texture between the area around this position and the central area, represents the texture value at position in the image, represents the texture value at the center of the local area, is a parameter for adjusting the sensitivity of texture change, usually with a value of 2; In this way, the degree of texture change in each local area is calculated, the spatial texture features of the image are extracted from the global context, and the change of local texture is quantified, thus providing a basis for subsequent blur judgment.

[0031] After extracting the local texture features, the degree of image chaos is quantified by calculating the local texture change degree of each local area, and a spatial texture chaos reference value is generated based on the quantification result of the local texture change degree to evaluate the overall texture disorder of the image. The generation expression of the spatial texture chaos reference value is: , where: is the quantified spatial texture chaos reference value, is the local texture feature at position in the image, is the global texture average feature of the image, is a parameter for controlling the perception of texture difference, usually selected as 3, k is the index of the currently processed local area. As the calculation process progresses, k will traverse all local areas, N is the total number of local areas divided in the image; By calculating the texture differences of all local areas, the spatial texture chaos reference value of the image is obtained. The spatial texture chaos reference value represents the overall texture disorder degree of the image. As the motion blur increases, the differences of local textures will become more blurred and chaotic, resulting in an increase in the spatial texture chaos reference value. In this way, the spatial texture chaos reference value can effectively quantify the image distortion caused by motion blur and provide a basis for subsequent blur judgment.

[0032] As can be seen from the spatial texture chaos reference value, the larger the value of the spatial texture chaos reference value generated after feature engineering processing of the distribution characteristics of the extracted image spatial texture, usually means the higher the blur degree of the image. The spatial texture chaos reference value of the image reflects the disorder degree of the texture distribution. When the image is blurred, especially due to motion blur, the details in the image will be lost, and the regularity and structure of the texture become blurred, resulting in more irregular and chaotic changes in local textures. Therefore, the spatial texture chaos reference value will increase significantly, indicating a higher degree of texture disorder in the image. On the contrary, when the image is clear, the texture distribution is more orderly, the details in the image are clearly visible, the changes in local textures are smoother and more regular, and the spatial texture chaos reference value will be lower. Therefore, the spatial texture chaos reference value can be used as an effective index to measure the image blur degree. A larger value indicates that the image is blurred, while a smaller value indicates that the image is clearer.

[0033] The blur prediction module inputs the blurred features after feature engineering into a pre-trained deep learning model, and predicts the blur degree of the current image through the deep learning model; Input the edge gradient reference value and spatial texture chaos reference value after feature engineering into a pre-trained deep learning model, generate a blur quantization coefficient based on the deep learning model, and predict the blur degree of the current image through the blur quantization coefficient.

[0034] The pre-trained deep learning model refers to a model that has been trained through a large number of labeled datasets and specific tasks before actual prediction. In the scenario of image blur prediction, the deep learning model usually learns through supervised learning using labeled training data to learn how to judge the blur degree of an image based on the features of the image (such as edge gradient reference value, spatial texture chaos reference value, etc.). The model continuously adjusts its internal parameters through backpropagation and optimization algorithms (such as gradient descent method) until it can efficiently establish an accurate mapping relationship between the input features and the actual blur degree of the image. During the training process, the model processes various types of images, identifies and extracts the key features affecting the image blur degree, and gradually learns how to fuse these features in order to effectively evaluate the blur degree of the image in future predictions.

[0035] This "pre-trained" process is usually completed in the development stage and trained using a large-scale and diverse image dataset. These datasets include images with various blur degrees, as well as corresponding true blur degree labels, which can be manually labeled or quantitative values obtained through other reliable methods (such as pixel offset of motion blur). Through multiple iterative trainings, the deep learning model can gradually improve its prediction accuracy and accurately infer the blur degree of the image from the input image features (such as edge gradient, spatial texture, etc.). When the training process is completed, the model can make real-time predictions based on the input of new image features. The trained deep learning model is essentially an optimized black-box system that can effectively process and infer unknown data based on the knowledge learned during training.

[0036] In the application scenario of image blur prediction, the edge gradient reference value and the spatial texture disorder reference value obtained after feature engineering processing are used as inputs and passed to a deep learning model. This model calculates an output value through forward propagation, which is usually called the blur quantization coefficient. This blur quantization coefficient is the quantization result of the current image blur and represents the degree of image blur. Through prior training, the deep learning model has learned how to extract the key information of image blur from these feature values and can even identify subtle blur signs from complex image feature combinations. In the training stage, the model not only learns how to use basic features such as edge gradients and texture disorder to judge image blur, but also masters the complex relationships between different image types and blur degrees through learning a large number of samples.

[0037] Specifically, the model may adopt a convolutional neural network (CNN) or other deep learning architectures suitable for processing image data. These models usually extract the spatial features of the image through multiple levels of convolutional operations, model the non-linear relationships using activation functions, and integrate the information of different levels through fully connected layers. During the training process, the network continuously adjusts the weights and biases to reduce the error between the predicted value and the true blur degree. After training, the model can quickly and accurately predict the blur degree when receiving new image features. Especially based on the edge gradient reference value and the spatial texture disorder reference value, the model can not only identify the significant features of blur, but also identify the blur type of the image (such as motion blur, focus blur, etc.) according to the patterns in the historical training data, so as to obtain a more accurate blur quantization coefficient.

[0038] This trained deep learning model has a high adaptability in practical applications and can handle different environments and different types of image blur situations. For actual detection tasks, users only need to input the edge gradient reference value and the spatial texture disorder reference value of the new image into the model, and the model can generate prediction results based on the existing knowledge. This deep learning-based prediction method is highly automated and intelligent, can accurately predict image blur without manual intervention, and shows strong robustness in various blur situations.

[0039] The deep learning model is not specifically limited here, as long as it can realize the comprehensive analysis of the edge gradient reference value and the spatial texture disorder reference value to generate the blur quantization coefficient is acceptable. To implement the technical solution of the present invention, the present invention provides a specific implementation method; the expression for generating the blur quantization coefficient is:

[0040] , where and are respectively the reference values of edge gradient and the reference value of spatial texture disorder of the preset proportionality coefficients, and and are both greater than 0. The preset proportionality coefficient refers to a constant coefficient set in advance during the model construction process, used to adjust the contribution ratio of the reference value of edge gradient and the reference value of spatial texture disorder to the fuzzy quantization coefficient . These preset proportionality coefficients are usually determined through experience or preliminary experiments, aiming to balance the influence of these two features on the final fuzzy quantization coefficient. Specifically, and represent the weighted coefficients of edge gradient and spatial texture, controlling their relative importance in the prediction model. The purpose of setting these coefficients is to ensure that the model can appropriately adjust its calculation process according to different image features. Usually, during the model training stage, these coefficients will be further adjusted through optimization algorithms to make them most suitable for the blur prediction of actual images.

[0041] It can be seen from the fuzzy quantization coefficient that the larger the performance value of the reference value of edge gradient generated after feature engineering processing of the gradient change feature of the extracted image edge, and the smaller the performance value of the reference value of spatial texture disorder generated after feature engineering processing of the distribution feature of the extracted image spatial texture, that is, the smaller the performance value of the fuzzy quantization coefficient generated when predicting the blur degree of the current image through a pre-trained deep learning model, indicating that the blur degree of the image is lower, and vice versa, indicating that the blur degree of the image is higher.

[0042] The image classification module divides the captured image into a normal image and a blurred image according to the prediction result of the deep learning model; Compare and analyze the fuzzy quantization coefficient generated when predicting the blur degree of the current image through a pre-trained deep learning model with the preset fuzzy quantization coefficient reference threshold, and divide the captured image. The specific division steps are as follows: If the generated fuzzy quantization coefficient is greater than the fuzzy quantization coefficient reference threshold, the captured image is divided into a blurred image; if the generated fuzzy quantization coefficient is less than or equal to the fuzzy quantization coefficient reference threshold, the captured image is divided into a normal image.

[0043] The normal image processing module, for the normal image, continues to keep the initial exposure time to expose when taking pictures with the micro camera, and performs subsequent tick feature extraction, identification, and positioning; The blurred image exposure adjustment module, for blurred images, adaptively regulates the initial exposure time according to the predicted result of the image blur degree by using the fuzzy logic algorithm, shortens the actual exposure duration during the micro-camera photographing, reduces the blur effect caused by movement, and performs subsequent tick feature extraction, recognition, and positioning after the image is determined to be a normal image; For normal images, continue to maintain the initial exposure time for photographing and perform subsequent tick feature extraction, recognition, and positioning. The purpose is to ensure clear image quality under appropriate exposure time conditions, so as to provide accurate image information for tick detection. A normal image means that the image is not affected by blur, with clear details, obvious edges and textures, which provides a reliable basis for tick feature extraction. By performing tick feature extraction on such images, features such as the shape, size, and position of ticks can be accurately identified, and the position of ticks can be further located and subsequent processing can be carried out. This process can ensure that the detection system efficiently and accurately identifies ticks, avoid misjudgment and missed judgment, improve the accuracy and timeliness of detection, and thus effectively prevent the risk of tick transmission.

[0044] For blurred images, adaptively regulate the initial exposure time according to the predicted result of the image blur degree by using the fuzzy logic algorithm. The specific steps are as follows: When processing blurred images, first calculate the blur quantization coefficient of the image through a deep learning model . The blur quantization coefficient quantifies the blur degree of the image, and the larger it is, the higher the image blur degree. In order to effectively adjust the exposure of the image, compare the blur quantization coefficient with the preset blur quantization coefficient reference threshold . Based on the fuzzy logic algorithm, calculate the exposure adjustment factor according to the difference between the two. In this process, adopt fuzzy inference rules, such as "the larger the blur quantization coefficient, the shorter the exposure time", and use fuzzy operations to deduce the size of the adjustment factor. The specific exposure adjustment factor calculation formula is: , where is the gain coefficient of fuzzy inference, which is used to adjust the sensitivity of exposure time adjustment; The function of this step is to quantify the degree of image blur according to the difference between the blur quantization coefficient and the blur quantization coefficient reference threshold , and generate the corresponding exposure adjustment factor through fuzzy inference. If the image blur degree is high, the blur quantization coefficient will be greater than the blur quantization coefficient reference threshold , and at this time the exposure adjustment factor increases, and the system will shorten the exposure time to reduce blur.

[0045] Based on the generated exposure adjustment factor Adjust the initial exposure time to optimize the shooting quality. The specific adjustment expression is: , where is the adjusted exposure time, that is, the actually used exposure time, is the initial exposure time, is the exposure adjustment coefficient, which controls the amplitude of the exposure time change. If the image blur degree is relatively high, the value of the exposure adjustment factor is relatively large, so that the adjusted exposure time will be significantly reduced; This step reduces the impact caused by motion blur by adaptively adjusting the exposure time, ensuring that clear images can still be obtained in the case of fast movement or blur. This adjustment guarantees the quality of the image, making feature extraction more accurate, especially in the subsequent tick recognition and positioning processes, improving the usability and accuracy of the image.

[0046] Reduce the image distortion caused by motion blur by adaptively adjusting the exposure time of the micro camera. When the image is recognized as a blurred image, the system automatically shortens the exposure time according to the predicted result of the image blur degree through the fuzzy logic algorithm, thereby reducing the blur effect caused by too long exposure and ensuring that the image clarity is optimized. This adjustment ensures that when shooting fast-moving objects or in an unstable environment, the details of the image can be presented more clearly, avoiding the influence of motion blur on feature extraction. Only when the image is confirmed to be clear and has a low blur degree, will subsequent tick feature extraction, recognition and positioning be carried out, ensuring the accuracy and reliability of tick detection and avoiding misjudgment or missed judgment caused by blurred images.

[0047] Through the above scheme, the problem of image blur caused by the fast movement of ticks can be effectively solved, and the accuracy of the detection system can be significantly improved. First, through the deep learning model combined with image processing technology, the image blur degree can be accurately predicted, and normal images and blurred images can be accurately identified and classified. For blurred images, the adaptive exposure time control method is adopted to reduce the motion blur effect caused by long-time exposure, thereby ensuring the image clarity and improving the accuracy of feature extraction. This optimization process reduces the occurrence of misjudgment, ensures that ticks can be detected and removed in time, and effectively reduces the risk of host infection by pathogens. At the same time, by dynamically adjusting the exposure time, the system can work stably in various environments, enhancing the robustness and reliability of the device in practical applications.

[0048] As Figure 2 shown: The working principle of this image-driven tick detection comb is to detect target objects (such as ticks) through a micro camera. When the device is started and the switch button is pressed, the micro camera starts to capture images of the target area and displays relevant information on the display screen. If the device detects the presence of a target object, the alarm indicator light will turn on and be accompanied by a sound warning. The handle part of the image-driven tick detection comb contains a collection tube for collecting ticks or other small objects, which can adsorb or collect ticks and other objects into the tube through the bristles. During use, the user can timely understand the working status of the device and the detected results according to the prompts on the display screen. The overall working process involves the coordination of multiple functions such as image capture, target recognition, alarm prompt, and object collection to achieve the purpose of accurately detecting and removing ticks.

[0049] The above formulas are all dimensionless and take their numerical values for calculation. The formula is obtained by collecting a large amount of data for software simulation to get a formula closest to the actual situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0050] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should all be covered within the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claimed rights.

[0051] Only some exemplary embodiments of the present invention have been described above by way of illustration. Undoubtedly, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and description are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

Claims

1. An image-driven tick detection comb integrated with a deep learning model, characterized in that, It includes an initial exposure time configuration module, an image data acquisition module, a feature extraction and blur quantization module, a blur degree prediction module, an image classification module, a normal image processing module, and a blurred image exposure adjustment module; The initial exposure time configuration module first determines the initial exposure time suitable for the current environment and detection requirements, and configures the parameters of the micro camera; The image data acquisition module, after taking a photo, comprehensively obtains various information of the image, providing complete data support for subsequent feature extraction; The feature extraction and blur quantization module, after obtaining the complete image data, extracts the feature information related to image blur through image processing technology, performs feature engineering processing on the extracted feature information, and preliminarily quantifies the image distortion caused by motion blur, providing a basis for subsequent blur degree judgment; The blur degree prediction module inputs the blurred features after feature engineering processing into a pre-trained deep learning model, and predicts the blur degree of the current image through the deep learning model; The image classification module divides the captured image into a normal image and a blurred image according to the result predicted by the deep learning model; The normal image processing module, for the normal image, continues to keep the initial exposure time to expose the micro camera when taking a photo, and performs subsequent tick feature extraction, identification, and positioning; The blurred image exposure adjustment module, for the blurred image, adaptively adjusts the initial exposure time according to the predicted result of the image blur degree using a fuzzy logic algorithm, shortens the actual exposure duration when the micro camera takes a photo, reduces the blur effect caused by motion, and performs subsequent tick feature extraction, identification, and positioning after the image is determined to be a normal image; 2. The image-driven tick detection comb integrating a deep learning model according to claim 1, characterized in that, Determine the initial exposure time suitable for the current environment and detection requirements based on static detection requirements. The static detection requirements refer to that when detecting ticks, it is preset that the ticks and their surrounding environment are in a relatively static state to ensure that the micro camera can capture clear and motion-blur-free images.

3. The image-driven tick detection comb integrating a deep learning model according to claim 1, characterized in that Extract the feature information related to image blur through image processing technology. Among them, the extracted feature information includes the gradient change of the image edge and the distribution of the image spatial texture. Perform feature engineering processing on the extracted feature information to generate an edge gradient reference value and a spatial texture chaos reference value respectively, and preliminarily quantify the image distortion caused by motion blur through the edge gradient reference value and the spatial texture chaos reference value, providing a basis for subsequent blur degree judgment.

4. The image-driven tick detection comb integrating a deep learning model according to claim 3, characterized in that, Input the edge gradient reference value and the spatial texture chaos reference value after feature engineering processing into a pre-trained deep learning model, generate a blur quantization coefficient based on the deep learning model, and predict the blur degree of the current image through the blur quantization coefficient.

5. The image-driven tick detection comb integrating a deep learning model according to claim 4, characterized in that, Compare and analyze the blur quantization coefficient generated when predicting the blur degree of the current image through a pre-trained deep learning model with a pre-set blur quantization coefficient reference threshold, and divide the captured image. The specific division steps are as follows: If the generated blur quantization coefficient is greater than the blur quantization coefficient reference threshold, then divide the captured image into a blurred image; If the generated fuzzy quantization coefficient is less than or equal to the fuzzy quantization coefficient reference threshold, the captured image is divided into a normal image.

6. The image-driven tick detection comb integrating a deep learning model according to claim 3, wherein, The specific steps for performing feature engineering on the gradient change features of the extracted image edges to generate an edge gradient reference value are as follows: The second derivative of the Laplacian operator is used to extract the gradient information of each point in the image, calculate the gradient of the image, capture the local brightness change of each pixel point in the image, and the extraction expression is as follows: , where: represents the image at the position The result of the Laplacian operator, which represents the change in the rate of change of pixel point brightness, can describe the degree of curvature of the image in the horizontal and vertical directions, and thus capture the edge information of the image, represents the pixel value of the image at the position , and represent the second derivatives in the horizontal and vertical directions respectively; Based on the gradient change of each pixel in the image, an edge gradient reference value is defined to quantify the blurriness of the image. The edge gradient reference value reflects the overall blurriness of the image by accumulating the intensity of the gradient change and using an exponential decay function. The specific calculation expression is as follows: , where represents the edge gradient reference value, represents the set of all pixel points in the image, is the absolute value of the gradient.

7. The image-driven tick detection comb integrating a deep learning model according to claim 3, wherein, The specific steps for performing feature engineering on the distribution features of the extracted image spatial texture to generate a spatial texture disorder reference value are as follows: The image is decomposed into multiple frequency sub-bands using wavelet transform, and different levels of local texture features are obtained from the decomposed frequency sub-bands. The texture information of each local region is represented as a feature vector, which reflects the local texture pattern of that region. To further quantify the disorder of these local texture features, a new local feature calculation formula is adopted, and the calculation expression is as follows: , where: represents the local texture change degree at position in the image, which reflects the difference in texture between the surrounding area and the central area at that position, represents the texture value at position in the image, represents the texture value at the center of the local region, is a parameter for adjusting the sensitivity of texture change; After the extraction of local texture features is completed, the degree of image chaos is quantified by calculating the local texture variation of each local region, and a spatial texture chaos reference value is generated based on the quantification result of the local texture variation to evaluate the overall texture disorder of the image. The generation expression of the spatial texture chaos reference value is: , where: is the quantified spatial texture chaos reference value, is the local texture feature at position in the image, is the global texture average feature of the image, is the parameter that controls the perception of texture differences, k is the index of the current processed local region. As the calculation process progresses, k will traverse all local regions, N is the total number of local regions divided in the image.

8. The image-driven tick detection comb integrating a deep learning model according to claim 5, characterized in that, For a blurred image, according to the predicted result of the image blur degree, an adaptive regulation of the initial exposure time is performed using a fuzzy logic algorithm. The specific steps are as follows: When processing a blurred image, first calculate the blur quantization coefficient of the image through a deep learning model , and quantify the blur degree of the image. In order to perform effective exposure adjustment on the image, compare the blur quantization coefficient with a preset reference threshold of the blur quantization coefficient , and based on the fuzzy logic algorithm, calculate the exposure adjustment factor according to the difference between the two . The specific calculation formula for the exposure adjustment factor is: , where is the gain coefficient of fuzzy inference, which is used to adjust the sensitivity of exposure time adjustment; Based on the generated exposure adjustment factor Adjust the initial exposure time to optimize the shooting quality. The specific adjustment expression is: , where is the adjusted exposure time, that is, the actual exposure time used, is the initial exposure time, is the exposure adjustment coefficient, which controls the amplitude of the exposure time change.

Images