Method, device and equipment for detecting and identifying tender shoots of tea leaves and storage medium
Through deep reinforcement learning network training of tea bud detection and recognition model, the problem of slow tea bud recognition speed and low accuracy is solved, and more efficient tea bud recognition is achieved.
Patent Information
- Application Number
- CN202411483345.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-23
- Publication Date
- 2025-01-10
AI Technical Summary
In the prior art, when identifying tea tender buds, the tea leaves have diverse growth postures, random distribution of tender buds, and mutual occlusion, and the influence of light, wind and rainfall factors at different times, the recognition speed is slow and the recognition accuracy is low.
The tea bud detection and recognition model is trained using deep reinforcement learning network. By collecting tea bud images, the variety and quality thresholds are determined, the image data set is screened, and the noise processing and positioning recognition is performed to reduce the influence of environmental factors.
It improves the recognition speed and accuracy of tea buds, and reduces the impact of tea growth posture and environmental factors on recognition.
Smart Images
Figure CN119295944A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of visual recognition technology, and in particular to a method, device, equipment and storage medium for detecting and identifying young tea buds. Background Art
[0002] With the rise of "Internet + Agriculture", intelligent technology has also become a new production method in the agricultural field. my country has a long history of tea culture development, and the tea economy plays an extremely important role in promoting regional economic construction. However, the cost of tea production is constantly rising, and the main cost link comes from tea picking. In the history of tea development, people have been constantly exploring how to pick tea efficiently.
[0003] Target recognition and detection is a hot topic in computer vision and digital image processing, and is widely used in robot navigation, intelligent video surveillance, industrial inspection, aerospace and many other fields. Reducing the consumption of human capital through computer vision has important practical significance. At present, the mainstream artificial intelligence technology to achieve automatic tea picking is mostly controlled by industrial computers due to stability and cost considerations.
[0004] However, due to the various growth postures of tea leaves, the random distribution and mutual occlusion of young buds, as well as the light, wind and rainfall factors brought about by different weather conditions at different times all affect the recognition and detection of young buds. The existing artificial intelligence technology has the problems of slow recognition speed and low recognition accuracy when identifying young tea buds. Summary of the invention
[0005] In view of this, it is necessary to provide a method, device, equipment and storage medium for detecting and identifying tea buds, so as to solve the problems of slow recognition speed and low recognition accuracy when identifying tea buds due to the various growth postures of tea leaves, the random distribution and mutual occlusion of the buds, and the light, wind and rainfall factors brought about by different weather conditions at different times.
[0006] In order to achieve the above technical objectives, this application adopts the following technical solutions: In a first aspect, the present application provides a method for detecting and identifying tea buds, comprising: Collecting images of tea buds to be identified and determining tea varieties based on the images of tea buds to be identified; Determine a tea bud image dataset that meets a preset tea bud quality threshold according to the tea bud image to be identified and the tea variety; Based on the tea bud image dataset, a deep reinforcement learning network training was performed to obtain a tea bud detection and recognition model. The tea bud image dataset is input into the tea bud detection and recognition model to identify the tea buds and determine the location information of the tea buds.
[0007] In a possible implementation, determining a tea bud image dataset that meets a preset tea bud quality threshold according to the tea bud image to be identified and the tea variety includes: Determine the density of tea buds according to the pixel conditions of the tea bud image to be identified; Classify and evaluate tea varieties to determine their grades; The tea bud evaluation value is calculated based on the tea bud density and tea variety grade; The tea bud image dataset is obtained according to the tea bud evaluation value and the preset tea bud quality threshold.
[0008] In a possible implementation, the tea bud evaluation value is calculated according to the tea bud density and the tea variety grade, including: Establish an initial tea bud quality evaluation model; The initial tea bud quality evaluation model is trained to obtain the target tea bud quality evaluation model; The tea bud density and tea variety grade are input into the target tea bud quality evaluation model to obtain the tea bud evaluation value.
[0009] In a possible implementation, a tea bud detection and recognition model is obtained by performing deep reinforcement learning network training based on a tea bud image dataset, including: The tea bud image dataset is input into the deep reinforcement learning network to determine the actual value of each control parameter; Modify the deep reinforcement learning network based on the actual values of each control parameter; Determine whether the preset termination condition is met. If so, output the current deep reinforcement learning network model as the tea bud detection and recognition model. Otherwise, use the next tea bud image dataset to train the deep reinforcement learning network.
[0010] In one possible implementation, the deep reinforcement learning network is modified based on the actual values of the control parameters, including: Calculate the loss value of tea bud detection and identification according to the actual value of each control parameter; Feedback correction is performed on the deep reinforcement learning network based on the loss value.
[0011] In a possible implementation, the tea bud image dataset is input into the tea bud detection and recognition model to recognize the tea buds and determine the location information of the tea buds, including: The tea bud image dataset is denoised to obtain a denoised tea bud image dataset; Based on the tea bud detection and recognition model, tea bud targets are identified from the denoised tea bud image dataset; The tea buds target is located based on visual recognition and coordinate transformation to obtain the position information of the tea buds.
[0012] In a possible implementation, performing denoising on a tea bud image dataset to obtain a denoised tea bud image dataset includes: Determine the data blocks where multiple noise points in the tea bud image data set are located; Assign denoising weights to data blocks based on the location of noise points; The pixel values after denoising of the noise points are calculated according to the denoising weights to obtain the denoised tea bud image dataset.
[0013] In a second aspect, the present application also provides a detection and identification device for tea buds, comprising: A variety identification module is used to collect images of tea buds to be identified and determine the tea variety based on the images of tea buds to be identified; A quality screening module, used to determine a tea bud image dataset that meets a preset tea bud quality threshold according to the tea bud images to be identified and the tea varieties; A model training module is used to perform deep reinforcement learning network training based on a tea bud image dataset to obtain a tea bud detection and recognition model; The identification and positioning module is used to input the tea bud image data set into the tea bud detection and recognition model to identify the tea buds and determine the location information of the tea buds.
[0014] In a third aspect, the present application also provides a tea bud detection and identification device, including a memory and a processor, wherein: Memory, used to store programs; The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps in the method for detecting and identifying young tea buds in any of the above implementations.
[0015] In a fourth aspect, the present application also provides a computer-readable storage medium for storing computer-readable programs or instructions, which, when executed by a processor, can implement the steps in the method for detecting and identifying tea buds in any of the above-mentioned implementations.
[0016] The beneficial effects of adopting the above embodiments are as follows: the present application relates to a method, device, equipment and storage medium for detecting and identifying tea buds, the method comprising: collecting tea bud images to be identified and determining tea varieties according to the tea bud images to be identified; determining tea bud image datasets that meet a preset tea bud quality threshold according to the tea bud images to be identified and the tea varieties; performing deep learning network training based on the tea bud image dataset to obtain a tea bud detection and identification model; inputting the tea bud image dataset into the tea bud detection and identification model to identify tea buds, and determining the location information of the tea buds. The present application first determines the tea variety according to the tea bud images to be identified, screens out tea bud image datasets that meet the quality of tea buds, and performs deep learning network training with the tea bud image dataset, thereby reducing the influence of different tea growth postures, random distribution of buds, mutual occlusion, and light factors, wind factors, and rainfall factors at different times and weather on the identification of tea buds, and improving the identification speed and accuracy of tea buds. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 A schematic diagram of a flow chart of an embodiment of a method for detecting and identifying young tea buds provided in the present application; Figure 2 Provided for this application Figure 1 A schematic diagram of a flow chart of an embodiment of step S102; Figure 3 Provided for this application Figure 2 A schematic diagram of a flow chart of an embodiment of step S203; Figure 4 Provided for this application Figure 1 A schematic flow chart of an embodiment of step S103; Figure 5 Provided for this application Figure 1 A schematic flow chart of an embodiment of step S104; Figure 6 Provided for this application Figure 5 A flowchart of an embodiment of step S501; Figure 7 A schematic structural diagram of an embodiment of a device for detecting and identifying tea buds provided in the present application; Figure 8 A schematic diagram of the structure of a tea bud detection and identification device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0018] The preferred embodiments of the present application are described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of the present application and are used together with the embodiments of the present application to illustrate the principles of the present application, and are not used to limit the scope of the present application.
[0019] In the description of the present application, “plurality” means two or more than two, unless otherwise clearly and specifically defined.
[0020] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.
[0021] The present application provides a method, device, equipment and storage medium for detecting and identifying young tea buds, which are described below respectively.
[0022] See also Figure 1 , Figure 1 A flow chart of an embodiment of a method for detecting and identifying young tea buds provided by the present application is provided. A specific embodiment of the present application discloses a method for detecting and identifying young tea buds, comprising: S101, collecting images of tea buds to be identified and determining tea varieties according to the images of tea buds to be identified; S102, determining a tea bud image dataset that meets a preset tea bud quality threshold according to the tea bud images to be identified and the tea varieties; S103, performing deep reinforcement learning network training based on the tea bud image dataset to obtain a tea bud detection and recognition model; S104, inputting the tea bud image dataset into the tea bud detection and recognition model to perform tea bud recognition and determine the position information of the tea bud.
[0023] In the above embodiment, the image of the tea buds to be identified is collected by a binocular camera, and the image of the tea buds to be identified is analyzed to determine the current tea variety. Different tea varieties have different growth cycles, which will affect the identification of the tea buds. It can also be considered that the tea variety is directly set.
[0024] The purpose of evaluating the quality of tea buds is to screen the data set. Since the range of tea buds to be identified is large, the image data collected by the binocular camera is also large. There may be a lot of useless data in the image, which is not suitable for deep reinforcement learning network training. The tea bud image data set that meets the preset tea bud quality threshold is determined through screening.
[0025] The deep learning model for tea shoot detection uses the YOLOV8n-C2f-DWR network model. YOLO (you only look once) only needs to browse once to identify the category and location of objects in the image. YOLO is called a region-free method, which is also called a single-stage (1-stage) model.
[0026] The YOLOV8n-C2f-DWR network model is based on the yolov8n network model. First, a Dilation-wise Residual (DWR) module is introduced to extract features from the high-level network. The multi-branch structure is used to expand the receptive field, where each branch uses a hole depth convolution with a different hole rate. Then, a SimpleInverted Residual (SIR) module is specially designed to extract features from the bottom layer of the network. This module has only a tiny receptive field of 3×3, but uses an inverted bottleneck structure to expand the number of channels to ensure stronger feature extraction capabilities. Finally, based on the DWR and SIR modules, an encoder-decoder style network DWRSeg is constructed, in which the decoder adopts a simple FCN-like structure. The decoder directly upsamples the feature maps using strong semantic information from the last two stages, and then splices them with feature maps from lower stages (containing rich details) for final prediction.
[0027] The tea bud detection and recognition model obtained through training can accurately identify tea buds and determine the location information of tea buds. The tea bud image dataset is input into the tea bud detection and recognition model to output the recognition result.
[0028] Compared with the prior art, the present embodiment provides a method for detecting and identifying tea buds, the method comprising: collecting tea bud images to be identified and determining tea varieties according to the tea bud images to be identified; determining a tea bud image dataset that meets a preset tea bud quality threshold according to the tea bud images to be identified and the tea varieties; performing deep learning network training based on the tea bud image dataset to obtain a tea bud detection and identification model; inputting the tea bud image dataset into the tea bud detection and identification model to identify tea buds, and determining the location information of the tea buds. The present application first determines the tea variety according to the tea bud images to be identified, screens out a tea bud image dataset that meets the quality of tea buds, and performs deep learning network training with the tea bud image dataset, thereby reducing the influence of different tea growth postures, random distribution of buds, mutual occlusion, and light factors, wind factors, and rainfall factors at different times and weather on the identification of tea buds, and improving the identification speed and accuracy of tea buds.
[0029] See also Figure 2 , Figure 2 Provided for this application Figure 1 In some embodiments of the present application, a tea bud image dataset that meets a preset tea bud quality threshold is determined according to the tea bud image to be identified and the tea variety, including: S201, determining the density of tea buds according to the pixel conditions of the tea bud image to be identified; S202, classifying and evaluating tea varieties to determine tea variety grades; S203, calculating the tea bud evaluation value according to the tea bud density and the tea variety grade; S204, obtaining a tea bud image dataset by screening according to the tea bud evaluation value and a preset tea bud quality threshold.
[0030] In the above embodiment, the higher the density of tea buds, the more the area needs to be picked, and there is an obvious difference in pixels between tea buds and non-tea buds. Therefore, the density of tea buds can be determined by the pixel conditions of the tea bud image to be identified.
[0031] The quality of young tea buds is also related to the tea variety. By counting the young tea buds of different tea varieties, the quality of young tea buds of different tea varieties can be evaluated and the tea varieties can be divided into grades.
[0032] Different weights are assigned to the tea bud density and tea variety grade, and the tea bud density, tea variety grade and corresponding weights are comprehensively calculated to obtain the tea bud evaluation value, which is used to screen the tea bud images to be identified to obtain a tea bud image dataset that meets the preset tea bud quality threshold.
[0033] See also Figure 3 , Figure 3 Provided for this application Figure 2 In some embodiments of the present application, the tea bud evaluation value is calculated based on the tea bud density and the tea variety grade, including: S301, establishing an initial tea bud quality evaluation model; S302, training the initial tea bud quality evaluation model to obtain a target tea bud quality evaluation model; S303, inputting the tea bud density and tea variety grade into the target tea bud quality evaluation model to obtain a tea bud evaluation value.
[0034] In the above embodiment, the established initial tea bud quality evaluation model takes the tea bud image to be identified as input and the tea bud evaluation value as output, but the initial tea bud quality evaluation model cannot accurately evaluate the tea buds. Therefore, the target tea bud quality evaluation model is obtained by training the initial tea bud quality evaluation model to improve the accuracy of the tea bud evaluation.
[0035] It should be understood that the model structure of the initial tea bud quality evaluation model includes but is not limited to deep neural network models (Deep Neural Networks, DNN), recurrent neural network models (Recurrent Neural Networks, RNN), convolutional network models (Convolutional Neural Networks, CNN), deep generative models (Deep Generative Models, DGM), generative adversarial networks (Generative Adversarial Networks, GAN), long short-term memory network models (Long / short term memory, LSTM), support vector machines (Support vector machines, SVM), deep cross model (Deep Crossing), etc.
[0036] See also Figure 4 , Figure 4 Provided for this application Figure 1 In some embodiments of the present application, a tea bud detection and recognition model is obtained by performing deep reinforcement learning network training based on a tea bud image dataset, including: S401, inputting the tea bud image dataset into the deep reinforcement learning network to determine the actual value of each control parameter; S402, modifying the deep reinforcement learning network based on the actual values of each control parameter; S403, judging whether the preset termination condition is met, if so, outputting the current deep reinforcement learning network model as the tea bud detection and recognition model, otherwise using the next tea bud image data set to train the deep reinforcement learning network.
[0037] In the above embodiment, determining the actual value of each control parameter can be regarded as a multi-label classification problem, where the labels are optional values of all control parameters. Since there are no labeled training samples, reinforcement learning is used to train and obtain the required control parameters.
[0038] The preset termination condition can be to stop training when the maximum number of times is reached, output the deep reinforcement learning network model, and obtain the tea bud detection and recognition model. It can also be to calculate the recognition accuracy and compare it with the preset accuracy threshold.
[0039] In some embodiments of the present application, the deep reinforcement learning network is modified based on the actual value of each control parameter, including: Calculate the loss value of tea bud detection and identification according to the actual value of each control parameter; Feedback correction is performed on the deep reinforcement learning network based on the loss value.
[0040] In the above embodiment, the normalized function region softmax can be used to calculate the probability of the optional value of each control parameter, and then the actual parameter value of the corresponding control parameter is sampled and selected according to the probability. After the actual parameter value of each control parameter is calculated, the deep reinforcement learning network is corrected, the loss value is calculated, and the parameters of the deep reinforcement learning network are updated and back-propagated accordingly to realize the training of the deep reinforcement learning network.
[0041] See also Figure 5 , Figure 5 Provided for this application Figure 1 In some embodiments of the present application, the tea bud image dataset is input into the tea bud detection and recognition model to perform tea bud recognition and determine the location information of the tea bud, including: S501, performing denoising processing on a tea bud image dataset to obtain a denoised tea bud image dataset; S502, identifying tea bud targets from a denoised tea bud image dataset based on a tea bud detection and recognition model; S503: Positioning the tea buds target based on visual recognition and coordinate conversion to obtain position information of the tea buds.
[0042] In the above embodiment, the tea bud image dataset is denoised to reduce the noise in the digital image so as not to affect the recognition of the tea bud target in the tea bud detection and recognition model. The tea bud detection and recognition model is a model trained by a deep reinforcement learning network, and can identify the tea bud target based on the denoised tea bud image dataset.
[0043] After identifying the tea buds, in order to facilitate the picking of the tea buds, it is necessary to perform visual recognition and coordinate transformation on the tea buds in the image to locate the tea buds and accurately obtain the position information of the tea buds.
[0044] See also Figure 6 , Figure 6Provided for this application Figure 5 In some embodiments of the present application, performing denoising on a tea bud image dataset to obtain a denoised tea bud image dataset includes: S601, determining data blocks where a plurality of noise points in a tea bud image data set are located; S602, assigning denoising weights to data blocks according to the positions of noise points; S603, calculating pixel values of noise points after denoising according to denoising weights to obtain a denoised tea bud image dataset.
[0045] In the above embodiment, a model is first established to identify multiple noise points in the tea bud image data set, and the noise points are marked with boxes of a preset size to obtain a data block with a standard box where the noise point is located.
[0046] The noise point position is the pixel position of the noise point. Different weights are assigned to different pixels in the image. The weights can be assigned according to the actual situation of denoising a variety of images. In order to accurately assign denoising weights, a large amount of data can be statistically analyzed to assign the best denoising weights to the noise point positions.
[0047] An image denoising model can be established through a Markov random field (MRF) model or a non-local self-similarity (NSS) model, and the assigned denoising weights and noise point positions are input into the image denoising model for calculation to obtain a denoised tea bud image dataset.
[0048] In order to better implement the detection and identification method of tea buds in the embodiment of the present application, based on the detection and identification method of tea buds, correspondingly, please refer to Figure 7 , Figure 7 This is a schematic diagram of the structure of an embodiment of a device for detecting and identifying tea buds provided by the present application. The present application embodiment provides a device for detecting and identifying tea buds 700, including: Variety identification module 710, used to collect the image of tea buds to be identified and determine the tea variety according to the image of tea buds to be identified; A quality screening module 720 is used to determine a tea bud image dataset that meets a preset tea bud quality threshold according to the tea bud images to be identified and the tea varieties; A model training module 730 is used to perform deep reinforcement learning network training based on a tea bud image dataset to obtain a tea bud detection and recognition model; The identification and positioning module 740 is used to input the tea bud image data set into the tea bud detection and recognition model to identify the tea buds and determine the location information of the tea buds.
[0049] It should be noted here that: the device 700 provided in the above embodiment can implement the technical solutions described in the above method embodiments. The specific implementation principles of the above modules or units can refer to the corresponding contents in the above method embodiments, which will not be repeated here.
[0050] See also Figure 8 , Figure 8 A schematic diagram of the structure of a tea bud detection and identification device provided in an embodiment of the present application. Based on the above-mentioned tea bud detection and identification method, the present application also provides a tea bud detection and identification device, which can be a computing device such as a mobile terminal, a desktop computer, a notebook, a palm computer, and a server. The tea bud detection and identification device 800 includes a processor 810, a memory 820, and a display 830. Figure 8 Only some components of the tea bud detection and identification device are shown, but it should be understood that it is not required to implement all of the shown components, and more or fewer components may be implemented instead.
[0051] In some embodiments, the memory 820 may be an internal storage unit of the device 800 for detecting and identifying tea buds, such as a hard disk or memory of the device 800 for detecting and identifying tea buds. In other embodiments, the memory 820 may also be an external storage device of the device 800 for detecting and identifying tea buds, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the device 800 for detecting and identifying tea buds. Further, the memory 820 may also include both an internal storage unit and an external storage device of the device 800 for detecting and identifying tea buds. The memory 820 is used to store application software and various data installed in the device 800 for detecting and identifying tea buds, such as a program code of the device 800 for detecting and identifying tea buds. The memory 820 may also be used to temporarily store data that has been output or is to be output. In one embodiment, a tea bud detection and identification program 840 is stored in the memory 820, and the tea bud detection and identification program 840 can be executed by the processor 810, so as to implement the tea bud detection and identification method of each embodiment of the present application.
[0052] In some embodiments, the processor 810 may be a central processing unit (CPU), a microprocessor or other data processing chip, and is used to run program codes or process data stored in the memory 820, such as executing a method for detecting and identifying young tea buds.
[0053] In some embodiments, the display 830 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch device, etc. The display 830 is used to display information on the tea bud detection and identification device 800 and to display a visual user interface. The components 810-830 of the tea bud detection and identification device 800 communicate with each other via a system bus.
[0054] In one embodiment, when the processor 810 executes the tea bud detection and identification program 840 in the memory 820, the steps in the above tea bud detection and identification method are implemented.
[0055] This embodiment further provides a computer-readable storage medium, on which a program for detecting and identifying tea buds is stored. When the program for detecting and identifying tea buds is executed by a processor, the following steps are implemented: Collecting images of tea buds to be identified and determining tea varieties based on the images of tea buds to be identified; Determine a tea bud image dataset that meets a preset tea bud quality threshold according to the tea bud image to be identified and the tea variety; Based on the tea bud image dataset, a deep reinforcement learning network training was performed to obtain a tea bud detection and recognition model. The tea bud image dataset is input into the tea bud detection and recognition model to identify the tea buds and determine the location information of the tea buds.
[0056] In summary, the present embodiment provides a method, device, equipment and storage medium for detecting and identifying tea buds, the method comprising: collecting tea bud images to be identified and determining tea varieties according to the tea bud images to be identified; determining a tea bud image dataset that meets a preset tea bud quality threshold according to the tea bud images to be identified and the tea varieties; performing deep learning network training based on the tea bud image dataset to obtain a tea bud detection and identification model; inputting the tea bud image dataset into the tea bud detection and identification model to identify tea buds and determine the location information of the tea buds. The present application first determines the tea variety according to the tea bud images to be identified, screens out tea bud image datasets that meet the quality of tea buds, and performs deep learning network training with the tea bud image dataset, thereby reducing the influence of different tea growth postures, random distribution of buds, mutual occlusion, and light factors, wind factors, and rainfall factors at different times and weather on the identification of tea buds, and improving the identification speed and accuracy of tea buds.
[0057] The above is only a preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by any technician familiar with the technical field within the technical scope disclosed in the present application should be covered within the protection scope of the present application.
Claims
1. A method for detecting and identifying tea buds, characterized in that: include: Collecting images of tea buds to be identified and determining tea varieties based on the images of tea buds to be identified; Determine a tea bud image dataset that meets a preset tea bud quality threshold according to the tea bud image to be identified and the tea variety; Based on the tea bud image dataset, a deep reinforcement learning network training was performed to obtain a tea bud detection and recognition model. The tea bud image dataset is input into the tea bud detection and recognition model to identify the tea buds and determine the location information of the tea buds.
2. The method for detecting and identifying tea buds according to claim 1, characterized in that: A tea bud image dataset that meets a preset tea bud quality threshold is determined according to the tea bud image to be identified and the tea variety, including: Determine the density of tea buds according to the pixel conditions of the tea bud image to be identified; Classify and evaluate tea varieties to determine their grades; The tea bud evaluation value is calculated based on the tea bud density and tea variety grade; The tea bud image dataset is obtained according to the tea bud evaluation value and the preset tea bud quality threshold.
3. The method for detecting and identifying tea buds according to claim 2, characterized in that: The tea bud evaluation value is calculated based on the tea bud density and tea variety grade, including: Establish an initial tea bud quality evaluation model; The initial tea bud quality evaluation model is trained to obtain the target tea bud quality evaluation model; The tea bud density and tea variety grade are input into the target tea bud quality evaluation model to obtain the tea bud evaluation value.
4. The method for detecting and identifying tea buds according to claim 1, characterized in that: Based on the tea bud image dataset, a deep reinforcement learning network training was performed to obtain a tea bud detection and recognition model, including: The tea bud image dataset is input into the deep reinforcement learning network to determine the actual value of each control parameter; Modify the deep reinforcement learning network based on the actual values of each control parameter; Determine whether the preset termination condition is met. If so, output the current deep reinforcement learning network model as the tea bud detection and recognition model. Otherwise, use the next tea bud image dataset to train the deep reinforcement learning network.
5. The method for detecting and identifying tea buds according to claim 4, characterized in that: The deep reinforcement learning network is modified based on the actual values of each control parameter, including: Calculate the loss value of tea bud detection and identification according to the actual value of each control parameter; Feedback correction is performed on the deep reinforcement learning network based on the loss value.
6. The method for detecting and identifying tea buds according to claim 1, characterized in that: The tea bud image dataset is input into the tea bud detection and recognition model to identify the tea buds and determine the location information of the tea buds, including: The tea bud image dataset is denoised to obtain a denoised tea bud image dataset; Based on the tea bud detection and recognition model, tea bud targets are identified from the denoised tea bud image dataset; The tea buds target is located based on visual recognition and coordinate transformation to obtain the position information of the tea buds.
7. The method for detecting and identifying tea buds according to claim 6, characterized in that: The tea bud image dataset is denoised to obtain a denoised tea bud image dataset, including: Determine the data blocks where multiple noise points in the tea bud image data set are located; Assign denoising weights to data blocks based on the location of noise points; The pixel values after denoising of the noise points are calculated according to the denoising weights to obtain the denoised tea bud image dataset.
8. A device for detecting and identifying tea buds, characterized in that: include: A variety identification module is used to collect images of tea buds to be identified and determine the tea variety based on the images of tea buds to be identified; A quality screening module, used to determine a tea bud image dataset that meets a preset tea bud quality threshold according to the tea bud images to be identified and the tea varieties; A model training module is used to perform deep reinforcement learning network training based on a tea bud image dataset to obtain a tea bud detection and recognition model; The identification and positioning module is used to input the tea bud image data set into the tea bud detection and recognition model to identify the tea buds and determine the location information of the tea buds.
9. A device for detecting and identifying young tea buds, characterized in that: comprising a memory and a processor, wherein: Memory, used to store programs; The processor is coupled to the memory and is used to execute the program stored in the memory to implement the steps of the method for detecting and identifying tea buds in any one of claims 1 to 7.
10. A computer-readable storage medium, characterized in that: Used to store computer-readable programs or instructions, which, when executed by a processor, can implement the steps of the method for detecting and identifying tea buds in any one of claims 1 to 7.