Image recognition-based needle-shaped tea humidity detection method and system

Abstract

The method and system for detecting the humidity of needle-shaped tea based on image recognition disclosed in this application obtains the humidity detection results of the images of needle-shaped tea to be detected. When the number of images of the to-be-detected that fall within the range of the fixation processing value is greater than a first predetermined value, the corresponding humidity morphology is determined as the humidity morphology of the needle-shaped tea. Based on the morphology, clustering is performed on the humidity detection result library of the images of the to-be-detected needle-shaped tea to form a set of needle-shaped tea images corresponding to the humidity morphology. Data configuration filtering is performed to determine whether it belongs to a balanced or unbalanced configuration, and the fixation processing range value corresponding to the humidity morphology is determined. This solution only needs to obtain the current humidity detection result of the needle-shaped tea, and can directly determine the humidity morphology of the needle-shaped tea through the processing range value. It can be achieved without running a complex model, which is more efficient and accurate. Furthermore, it does not require the use of intermediate features such as color and brightness type, which reduces computational complexity and further reduces false recognition or missed recognition.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to a method and system for detecting the moisture content of needle-shaped tea leaves based on image recognition. Background Technology

[0002] Currently, the market supply of needle-shaped tea, especially needle-shaped green tea, is mainly of the light-aroma type. With the improvement of living standards and the diversification of demands, the need for floral-aroma needle-shaped tea is increasing, but traditional production processes cannot produce consistently high-quality floral-aroma needle-shaped tea. To meet market demand, this invention overcomes the shortcomings of the original production process and provides a method for processing floral-aroma needle-shaped tea using a temperature and humidity control system to produce consistently high-quality floral-aroma needle-shaped tea.

[0003] The raw materials for processing needle-shaped black tea vary greatly due to differences in climate, geographical conditions, varieties, and the harvesting season. To process these diverse raw materials into needle-shaped black tea with a uniform standard of quality and flavor, it is essential to regulate and control the fermentation process, including temperature, humidity, the moisture content of the raw tea leaves, and the brightness of the fermentation environment. In traditional needle-shaped black tea processing, this control relies entirely on the tea master's intuition and experience. Different tea masters have vastly different perceptions, and even apprentices taught by the same master, with identical instruction, produce significantly different needle-shaped black teas. Therefore, based on summarizing traditional experience, a key factor influencing the fermentation of needle-shaped black tea—humidity—has been identified. Regulating and controlling these factors ensures the consistency of needle-shaped black tea quality. This invention utilizes modern technology to automatically regulate fermentation temperature, humidity, and brightness to enhance enzyme activity and achieve rapid fermentation. Summary of the Invention

[0004] In view of the above problems, this application provides a method, device, equipment, storage medium, system, and needle-shaped tea based on image recognition for detecting the humidity of needle-shaped tea, which can accurately, efficiently, and in real time detect the humidity morphology of needle-shaped tea with low computational complexity.

[0005] To achieve the above objectives, the first aspect of this application provides a method for detecting the moisture content of needle-shaped tea leaves based on image recognition, comprising:

[0006] Obtain the humidity detection results of the image of the needle-shaped tea to be detected. The humidity detection results include at least two detection parameters.

[0007] If the number of images to be detected for a detection parameter that falls within the range of the fixation process is greater than a first predetermined value, the humidity morphology corresponding to the fixation process range value is determined as the humidity morphology of needle-shaped tea.

[0008] By acquiring humidity detection results including at least two detection parameters, and determining whether the number of images to be detected for the detection parameters falling within the range of the withering process is greater than a first predetermined value, the humidity morphology of needle-shaped tea can be directly determined. This is unaffected by the acquisition environment, and the data processing and transmission have low requirements for computing power and bandwidth. Furthermore, personalized humidity morphology detection can be achieved by using personalized withering process range values. Thus, the humidity morphology of specific needle-shaped tea can be detected accurately, efficiently, and in real time with low computational complexity, while improving the needle-shaped tea experience and reducing application costs.

[0009] As one possible implementation of the first aspect, at least two detection parameters are used to indicate moisture content and / or shape parameters. Thus, by using at least two indicators to indicate moisture content and / or shape parameters, the moisture condition of needle-shaped tea can be comprehensively and accurately perceived, facilitating the determination of its moisture morphology by comprehensively considering various moisture characteristics during needle-shaped tea processing, and further improving the accuracy of moisture morphology detection.

[0010] As one possible implementation of the first aspect, the humidity profile includes either a humidity profile to be treated or a dry humidity profile, and the blanching range value corresponds to either the humidity profile to be treated or the dry humidity profile. Therefore, by using the blanching range value corresponding to the humidity profile to be treated or the dry humidity profile, the humidity profile can be directly determined as either to be treated or dry, without relying on intermediate features such as color or brightness, thus reducing computational complexity and minimizing false or missed identifications.

[0011] As one possible implementation of the first aspect, the range value of the fixation treatment includes a treatment range value that corresponds one-to-one with at least two detection parameters. Therefore, by including the treatment range value corresponding to each detection parameter in the fixation treatment range value, it is easier to comprehensively consider various moisture characteristics during needle-shaped tea processing to determine its moisture morphology, thereby further improving the accuracy of moisture morphology detection.

[0012] As one possible implementation of the first aspect, the range value for the final processing is obtained in the following way:

[0013] Clustering is performed on the humidity detection result library of needle-shaped tea images to form a needle-shaped tea image set corresponding to humidity patterns. The humidity detection result library of needle-shaped tea images contains the humidity detection results of the needle-shaped tea images to be detected obtained in advance.

[0014] Data configuration filtering was performed on the needle-shaped tea image set to determine whether the data configuration of the needle-shaped tea image set belonged to a balanced configuration or an unbalanced configuration;

[0015] Based on the data configuration of the needle-shaped tea image set, the range of values ​​for fixation treatment corresponding to the humidity morphology is determined.

[0016] Therefore, by clustering the humidity detection results of needle-shaped tea, the range value of the fixation treatment for each needle-shaped tea can be obtained. The fixation treatment range value can be set according to the specific type of needle-shaped tea, thereby realizing the personalization of humidity morphology detection, further improving the accuracy of humidity morphology detection, and reducing the omission or misidentification caused by the uniformity of detection standards.

[0017] As one possible implementation of the first aspect, data configuration and filtering of the needle-shaped tea image set is performed, specifically including:

[0018] The humidity detection results in the needle-shaped tea image set are divided into pixel grids to obtain multiple pixel grids, and each pixel grid in the multiple pixel grids has the same number of pixels;

[0019] The pixel grid with the largest number of pixels among multiple pixel grids is used as the mean range of the needle-shaped tea image set;

[0020] When the number of pixels in the first range of the needle-shaped tea image set is equal to the number of pixels in the second range of the needle-shaped tea image set, the data configuration of the needle-shaped tea image set is a balanced configuration.

[0021] When the number of pixels in the first range of the needle-shaped tea image set is not equal to the number of pixels in the second range of the needle-shaped tea image set, the data configuration of the needle-shaped tea image set is an unbalanced configuration.

[0022] The first range and the second range are obtained by dividing the data range of the needle-shaped tea image set according to the mean range.

[0023] Therefore, by dividing the range into pixel grids and comparing the number of data points on the left and right halves of the mean range, it is possible to determine whether the needle-shaped tea image set is in a balanced or unbalanced configuration. This approach combines low complexity with accurate data configuration filtering, which can further reduce hardware costs and improve the accuracy of needle-shaped tea moisture morphology detection.

[0024] As one possible implementation of the first aspect, the image recognition-based needle-shaped tea humidity detection method further includes: when the number of images to be detected for the detection parameter falling within the range of the withering process value is less than or equal to a first predetermined value among at least two detection parameters of the humidity detection result, determining the humidity morphology of the needle-shaped tea based on the difference between the humidity detection result and the first needle-shaped tea reference humidity parameter value. Therefore, determining the humidity morphology by the difference between the humidity detection result and the first needle-shaped tea reference humidity parameter value can assist in detecting the humidity morphology of needle-shaped tea when the withering process range value cannot determine the humidity morphology, thereby achieving real-time, efficient, and accurate detection of the humidity morphology of needle-shaped tea with lower computational complexity.

[0025] As one possible implementation of the first aspect, the humidity profile includes a humidity profile to be processed or a dry humidity profile, and the first needle-shaped tea reference humidity parameter value corresponds to the humidity profile to be processed or the dry humidity profile. Thus, by using the first needle-shaped tea reference humidity parameter value corresponding to the humidity profile to be processed or the dry humidity profile, the humidity profile can be directly determined as either to be processed or dry, without relying on intermediate features such as color or brightness, reducing computational complexity and minimizing false or false identification.

[0026] As one possible implementation of the first aspect, the humidity profile of the needle-shaped tea is determined based on the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea, including:

[0027] Of the two reference humidity parameter values ​​for first-needle-shaped tea, the humidity pattern corresponding to the reference humidity parameter value for first-needle-shaped tea that is closer to the humidity detection result is determined as the humidity pattern of the needle-shaped tea; or,

[0028] Among three or more reference humidity parameter values ​​for first needle-shaped tea, the humidity pattern corresponding to the first needle-shaped tea reference humidity parameter value closest to the humidity detection result is determined as the humidity pattern of the needle-shaped tea.

[0029] By determining the first reference humidity parameter value of the needle-shaped tea that is closer to or closest to the humidity detection result, the humidity pattern can be directly determined, enabling real-time, efficient, and accurate detection of the humidity pattern of needle-shaped tea with low computational complexity.

[0030] As one possible implementation of the first aspect, the humidity profile of the needle-shaped tea is determined based on the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea. This includes: when the absolute value of the difference between the first difference and the second difference is greater than the average of the first difference and the second difference, the humidity profile is determined to be the humidity profile to be processed; when the absolute value of the difference between the first difference and the second difference is less than or equal to the average of the first difference and the second difference, the humidity profile is determined to be the dry humidity profile. The first difference is the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea corresponding to the humidity profile to be processed, and the second difference is the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea corresponding to the dry humidity profile. Therefore, by comparing the differences and the average of the differences, the reference humidity parameter value of the first needle-shaped tea that is closer to or closest to the humidity detection result is determined, which has low computational complexity and is easy to implement.

[0031] As one possible implementation of the first aspect, the image recognition-based method for detecting the humidity of needle-shaped tea further includes: when the humidity detection result of the needle-shaped tea image to be detected has a fall rate greater than or equal to a third predetermined value, adding the humidity detection results of the detection parameters that fall within the range of the withering process to the humidity detection result library of the needle-shaped tea image to be detected to a number of images to be detected that are less than or equal to a first predetermined value.

[0032] When the humidity detection result has a high fall rate, humidity detection results with a number of images falling within the range of the fixation process that are less than or equal to a first predetermined value are added to the humidity detection result library. This can improve the needle-shaped tea data images while reducing redundant data entering the humidity detection result library, further reducing the amount of data calculation and computational complexity.

[0033] As one possible implementation of the first aspect, the image recognition-based needle-shaped tea humidity detection method further includes: when the duration of the humidity mode being dry is greater than or equal to a second predetermined value, adjusting the humidity mode using one or more adjustment mechanisms including the following: initial kneading mode, re-kneading mode, second fermentation mode, and hot air dehydration mode.

[0034] Therefore, multiple adjustment mechanisms can be used to regulate the humidity state when the dryness and humidity state persists for a period of time, thus achieving timely and effective intervention in the drying and humidity state of needle-shaped tea.

[0035] As one possible implementation of the first aspect, the adjustment mechanism is determined based on the drying priority of the needle-shaped tea, which is determined according to the humidity detection results and the range of fixation treatment corresponding to the humidity morphology. Therefore, by determining the drying priority through humidity detection results and the fixation treatment range, and by determining the adjustment mechanism based on the drying priority, appropriate methods can be used to adjust the state for different degrees of dryness, significantly improving the effectiveness of humidity morphology intervention in drying.

[0036] As one possible implementation of the first aspect, the image recognition-based method for detecting the humidity of needle-shaped tea further includes: re-determining the first drying priority of the needle-shaped tea after a first time interval, wherein the first time interval corresponds to the first drying priority. Therefore, re-evaluating at different time intervals for different degrees of dryness better reflects the actual situation of improving the humidity morphology of needle-shaped tea and can significantly enhance the effectiveness of humidity morphology intervention.

[0037] As one possible implementation of the first aspect, the mode in the regulation mechanism is randomly selected. Therefore, by randomly selecting the mode in the regulation mechanism, the resilience of the regulation mechanism can be effectively improved.

[0038] A second aspect of this application provides a needle-shaped tea moisture detection system based on image recognition, comprising:

[0039] An acquisition device is used to acquire the humidity detection result of an image of a needle-shaped tea to be detected, the humidity detection result including at least two detection parameters;

[0040] A determining device is used to determine the humidity morphology corresponding to the fixation processing range value as the humidity morphology of needle-shaped tea when the number of images to be detected for a detection parameter falling within the fixation processing range value is greater than a first predetermined value among at least two detection parameters.

[0041] Therefore, by acquiring humidity detection results including at least two detection parameters, and determining whether the number of images to be detected for the detection parameters falling within the range of the withering process is greater than a first predetermined value, the humidity morphology of needle-shaped tea can be directly determined. This is unaffected by the acquisition environment, and the data processing and transmission have low requirements for computing power and bandwidth. Furthermore, personalized humidity morphology detection can be achieved by using personalized withering process range values. Thus, the humidity morphology of specific needle-shaped tea can be accurately, efficiently, and in real time detected with low computational complexity, while improving the needle-shaped tea experience and reducing application costs.

[0042] As one possible implementation of the second aspect, at least two detection parameters are used to indicate moisture content and / or shape parameters. Thus, by using at least two indicators to indicate moisture content and / or shape parameters, the moisture condition of needle-shaped tea can be comprehensively and accurately perceived, facilitating the determination of its moisture morphology by comprehensively considering various moisture characteristics during needle-shaped tea processing, and further improving the accuracy of moisture morphology detection.

[0043] As one possible implementation of the second aspect, the humidity profile includes a pending humidity profile and a dry humidity profile, with the blanching range value corresponding to either the pending or dry humidity profile. Therefore, by using the blanching range value corresponding to the pending or dry humidity profile, the humidity profile can be directly determined as pending or dry, without relying on intermediate features such as color or brightness, thus reducing computational complexity and minimizing false or missed identifications.

[0044] As a possible implementation of the second aspect, the image recognition-based needle-shaped tea moisture detection system further includes: a computing device, which is used for:

[0045] Clustering is performed on the humidity detection result library of needle-shaped tea images to form a needle-shaped tea image set corresponding to humidity patterns. The humidity detection result library of needle-shaped tea images contains the humidity detection results of the needle-shaped tea images to be detected obtained in advance.

[0046] Data configuration filtering was performed on the needle-shaped tea image set to determine whether the data configuration of the needle-shaped tea image set belonged to a balanced configuration or an unbalanced configuration;

[0047] Based on the data configuration of the needle-shaped tea image set, the range of values ​​for fixation treatment corresponding to the humidity morphology is determined.

[0048] Therefore, by clustering the humidity detection results of needle-shaped tea, the range value of the fixation treatment for each needle-shaped tea can be obtained. The fixation treatment range value can be set according to the specific type of needle-shaped tea, thereby realizing the personalization of humidity morphology detection, further improving the accuracy of humidity morphology detection, and reducing the omission or misidentification caused by the uniformity of detection standards.

[0049] As one possible implementation of the second aspect, the computing device is specifically used for:

[0050] The humidity detection results in the needle-shaped tea image set are divided into pixel grids to obtain multiple pixel grids, and each pixel grid in the multiple pixel grids has the same number of pixels;

[0051] The pixel grid with the largest number of pixels among multiple pixel grids is used as the mean range of the needle-shaped tea image set;

[0052] When the number of pixels in the first range of the needle-shaped tea image set is equal to the number of pixels in the second range of the needle-shaped tea image set, the data configuration of the needle-shaped tea image set is a balanced configuration.

[0053] When the number of pixels in the first range of the needle-shaped tea image set is not equal to the number of pixels in the second range of the needle-shaped tea image set, the data configuration of the needle-shaped tea image set is an unbalanced configuration.

[0054] The first range and the second range are obtained by dividing the data range of the needle-shaped tea image set according to the mean range.

[0055] Therefore, by dividing the range into pixel grids and comparing the number of data points on the left and right halves of the mean range, it is possible to determine whether the needle-shaped tea image set is in a balanced or unbalanced configuration. This approach combines low complexity with accurate data configuration filtering, which can further reduce hardware costs and improve the accuracy of needle-shaped tea moisture morphology detection.

[0056] As a possible implementation of the second aspect, the determining device is further configured to determine the humidity profile of the needle-shaped tea based on the difference between the humidity detection result and the first needle-shaped tea reference humidity parameter value when the number of images to be detected for the detection parameter falling within the range of the withering process is less than or equal to a first predetermined value among at least two detection parameters. Thus, determining the humidity profile by the difference between the humidity detection result and the first needle-shaped tea reference humidity parameter value can assist in detecting the humidity profile of the needle-shaped tea when the range of the withering process cannot determine the humidity profile, thereby achieving real-time, efficient, and accurate detection of the humidity profile of the needle-shaped tea with lower computational complexity.

[0057] As one possible implementation of the second aspect, the determining device is specifically used to determine the humidity pattern of the needle-shaped tea as the humidity pattern of the tea based on the humidity pattern corresponding to the first needle-shaped tea reference humidity parameter value that is closer to the humidity detection result among two first needle-shaped tea reference humidity parameter values; or, based on three or more first needle-shaped tea reference humidity parameter values, to determine the humidity pattern corresponding to the first needle-shaped tea reference humidity parameter value that is closest to the humidity detection result as the humidity pattern of the tea. Thus, by using the first needle-shaped tea reference humidity parameter value corresponding to the humidity pattern to be processed or the dry humidity pattern, the humidity pattern can be directly determined as either to be processed or dry, without relying on intermediate features such as color or brightness, reducing computational complexity and minimizing false or missed identifications.

[0058] As one possible implementation of the second aspect, the determining device is specifically used for:

[0059] When the absolute value of the difference between the first difference and the second difference is greater than the average of the first difference and the second difference, the humidity pattern is determined to be the humidity pattern to be processed.

[0060] When the absolute value of the difference between the first difference and the second difference is less than or equal to the average of the first difference and the second difference, the humidity pattern is determined to be a dry humidity pattern.

[0061] The first difference is the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea corresponding to the humidity state to be processed. The second difference is the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea corresponding to the dry humidity state.

[0062] Therefore, by comparing the differences and the mean between the differences, the first reference humidity parameter value of the needle-shaped tea that is closer to or closest to the humidity detection result can be determined. This method has low computational complexity and is easy to implement.

[0063] As a possible implementation of the second aspect, the image recognition-based needle-shaped tea humidity detection system further includes: a database update device, used to add humidity detection results whose number of images containing detection parameters falling within the range of the withering process is less than or equal to a first predetermined value to the humidity detection result database of the needle-shaped tea images when the humidity detection result inclusion rate of the images to be detected is greater than or equal to a third predetermined value. Thus, when the humidity detection result inclusion rate is high, adding humidity detection results whose number of images containing detection parameters falling within the range of the withering process is less than or equal to the first predetermined value to the humidity detection result database can improve the needle-shaped tea data images while reducing redundant data entering the humidity detection result database, further reducing the amount of data computation and computational complexity.

[0064] As a possible implementation of the second aspect, the image recognition-based needle-shaped tea humidity detection system further includes: an adjustment device, used to adjust the humidity state using one or more adjustment mechanisms including the following methods when the duration of the humidity state being dry is greater than or equal to a second predetermined value: initial rolling, re-rolling, second withering, and hot air dehydration. Thus, it is possible to adjust the humidity state using multiple adjustment mechanisms when the dry humidity state persists for a period of time, achieving timely and effective intervention in the drying humidity state of needle-shaped tea.

[0065] As one possible implementation of the second aspect, the adjustment mechanism is determined based on the drying priority of the needle-shaped tea. The drying priority is determined based on the humidity detection results and the range of fixation treatment corresponding to the humidity morphology. Therefore, by determining the drying priority through humidity detection results and the fixation treatment range, and by determining the adjustment mechanism based on the drying priority, appropriate methods can be used to adjust the state for different degrees of dryness, significantly improving the effectiveness of humidity morphology intervention in drying.

[0066] As a possible implementation of the second aspect, the regulating device is also used to redetermine the first drying priority of the needle-shaped tea after the first time interval, the first time interval corresponding to the first drying priority. Thus, different time intervals can be used to reassess for different degrees of dryness, which is more in line with the actual situation of improving the moisture morphology of needle-shaped tea and can significantly improve the effect of moisture morphology intervention.

[0067] As one possible implementation of the second aspect, the mode in the regulation mechanism is randomly selected by the regulating device. Therefore, by randomly selecting the mode in the regulation mechanism, the resistance of the regulation mechanism can be effectively improved.

[0068] A third aspect of this application provides a computing device, comprising: one or more processors and a memory, the memory storing program instructions, which, when executed by the one or more processors, cause the processors to perform the image recognition-based needle-shaped tea humidity detection method of the first aspect and possible implementations.

[0069] The aforementioned computing device can accurately, efficiently, and in real-time detect the moisture content of specific needle-shaped teas with low computational complexity, while improving the needle-shaped tea experience and reducing application costs.

[0070] The fourth aspect of this application provides a computer-readable storage medium having program instructions stored thereon, which, when executed by a computer, cause the computer to perform the image recognition-based needle-shaped tea humidity detection method of the first aspect and its possible implementations.

[0071] The fifth aspect of this application provides a needle-shaped tea monitoring system, including the image recognition-based needle-shaped tea humidity detection system of the second aspect and possible implementations, or the computing device of the third aspect.

[0072] The sixth aspect of this application provides a needle-shaped tea, including the image recognition-based needle-shaped tea humidity detection system of the second aspect and possible implementations, the computing device of the third aspect, or the needle-shaped tea monitoring system of the fifth aspect.

[0073] This application embodiment only needs to obtain the current humidity detection result of the needle-shaped tea to directly determine the humidity state of the needle-shaped tea through processing range values, which can be achieved without running a complex model, making it more efficient and accurate. Furthermore, it does not require intermediate features such as color and brightness type, reducing computational complexity and further reducing false or missed identifications. Attached Figure Description

[0074] The various features of the present invention and the relationships between them are further explained below with reference to the accompanying drawings. The drawings are exemplary; some features are not shown to scale, and some drawings may omit conventional features in the field of this application that are not essential to this application, or additional features that are not essential to this application may be shown. The combination of features shown in the drawings is not intended to limit the present application. Furthermore, throughout this specification, the same reference numerals refer to the same things. Specific descriptions of the drawings are as follows:

[0075] Figure 1 This is an exemplary flowchart of the image recognition-based moisture detection method for needle-shaped tea in this application embodiment;

[0076] Figure 2 This is an exemplary structural diagram of the image recognition-based needle-shaped tea moisture detection system provided in the embodiments of this application;

[0077] Figure 3 This is an exemplary flowchart of humidity morphology detection in the second embodiment of this application;

[0078] Figure 4 This is an exemplary flowchart of clustering in the embodiments of this application;

[0079] Figure 5 This is an exemplary flowchart of the data filtering configuration in the embodiments of this application;

[0080] Figure 6 This is an exemplary flowchart of the specific implementation of adjusting the humidity of needle-shaped tea in the embodiments of this application; Detailed Implementation

[0081] The terms "first," "second," "third," and similar terms used in the specification and claims are used only to distinguish similar objects and do not represent a specific ordering of objects. It is understood that, where permissible, a specific order or sequence may be interchanged so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0082] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments and corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0083] The term "comprising" as used in the specification and claims should not be construed as limiting itself to what follows; it does not exclude other components or steps. Therefore, it should be interpreted as specifying the presence of the mentioned feature, integral, step, or component, but does not exclude the presence or addition of one or more other features, integrals, steps, or components, or groups thereof. Thus, the statement "an apparatus comprising device A and device B" should not be limited to a device consisting solely of device A and device B.

[0084] The terms "some embodiments" or "embodiments" used in this specification mean that a particular feature, structure, or characteristic described in conjunction with that embodiment is included in at least some embodiments of the invention. Therefore, the terms "some embodiments" or "in embodiments" appearing throughout this specification do not necessarily refer to the same embodiment, but may refer to the same embodiment. Furthermore, in one or more embodiments, the particular features, structures, or characteristics can be combined in any suitable manner, as will be apparent to those skilled in the art from this disclosure.

[0085] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. In case of any inconsistency, the meaning set forth in this specification or derived from the content described herein shall prevail. Furthermore, the terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit the scope of this application.

[0086] In order to accurately describe the technical content of this application and to accurately understand the present invention, the following explanations or definitions of the terms used in this specification are given before describing the specific embodiments.

[0087] Supervised learning is a type of machine learning where the goal is to model the relationship between the features and labels of samples, and each sample in the training set has a label.

[0088] Support Vector Machine (SVM) is a binary classification algorithm and one of the methods in supervised learning. It is based on a linear function and only outputs the category.

[0089] Unsupervised learning refers to automatically learning the inherent properties and patterns of data from training samples that do not contain target labels.

[0090] Clustering divides a set of needle-shaped tea images into several typically disjoint subsets, each called a cluster. Through such a division, each cluster may correspond to some potential concepts or categories.

[0091] The mean is the middle number in a set of data arranged in order. It represents a value in a sample, population, or probability configuration that divides the set of values ​​into two equal parts.

[0092] Data configuration, also known as frequency configuration.

[0093] In a balanced configuration, most frequencies are concentrated in the central position, with the frequency configurations at both ends being roughly symmetrical, and the configuration curve is bell-shaped.

[0094] An unbalanced configuration has an asymmetrical frequency distribution, with the set mean biased to one side, resulting in an asymmetrical configuration curve.

[0095] The treatment range value represents the range of error within which a sample estimates the population mean. Selecting a treatment range value [a, b] guarantees that the result between a and b contains the population mean at a predetermined confidence level. For example, with a 95% confidence level, assuming 100 samples are taken, the treatment range value [a, b] will contain the population mean in 95 of those samples.

[0096] The mean range is the range of values ​​used to divide a set of needle-shaped tea images into two parts. If the data configuration of the needle-shaped tea image set is balanced, the two parts obtained by dividing the image set by the mean range have the same number of pixels. If the data configuration of the needle-shaped tea image set is unbalanced, the two parts obtained by dividing the image set by the mean range have different numbers of pixels.

[0097] The data range of the needle-shaped tea image set is the numerical range of the data in the needle-shaped tea image set. For example, after normalization, the numerical range of the data in the needle-shaped tea image set is (0, 1), then the data range of the needle-shaped tea image set can be (0, 1).

[0098] To address the aforementioned issues, this application provides a method, apparatus, device, and computer-readable storage medium for detecting the humidity of needle-shaped tea based on image recognition. The humidity profile of the needle-shaped tea can be determined by analyzing its humidity detection results and processing range. This method has low computational complexity, enabling accurate, efficient, and real-time detection of the humidity profile of specific needle-shaped teas with minimal computational complexity. It is applicable both offline and online, and suitable for various types of needle-shaped teas and various cabin environments. By accurately detecting the humidity profile of needle-shaped tea, it ensures accurate tea drinking while improving the tea-drinking experience and reducing application costs.

[0099] The technical solution of this application embodiment involves multiple weighings and image acquisitions of the specified needle-shaped tea during the moisture evaporation process to obtain several changes in mass and images of the specified needle-shaped tea as moisture evaporates. Weighing and image acquisition are performed simultaneously, with each data acquisition yielding a change in mass and a corresponding change image. Specifically, data acquisition can be performed on the specified needle-shaped tea at preset time intervals, such as every minute. The dried mass of the specified needle-shaped tea after moisture evaporation is obtained, representing the mass corresponding to when the moisture evaporation is minimal or even close to zero. By determining the actual humidity of the specified needle-shaped tea at each mass value, the initial image, each change image, and the corresponding actual humidity of the specified needle-shaped tea can be determined. Using the initial image, change images, and other images as input, and the corresponding actual humidity as output, as a training set, a model is learned and trained to obtain a trained humidity detection model for the specified needle-shaped tea. The humidity detection model can be a neural network model based on machine learning. During the moisture evaporation process of the specified needle-shaped tea, the difference between two adjacent mass values ​​can be determined for each acquired mass value (including the initial mass and several change masses mentioned above). If the difference is not greater than the preset threshold, it indicates that the mass value of the specified needle-shaped tea changes relatively quickly, meaning that the moisture evaporation thickness of the specified needle-shaped tea is also relatively fast. At this time, the moisture evaporation process of the needle-shaped tea is still in progress. If the difference is less than the preset threshold, it indicates that the mass value of the specified needle-shaped tea changes relatively slowly, and its moisture evaporation thickness is also slow. It can be considered that the moisture evaporation process of the needle-shaped tea has ended, and the smaller of the two mass values ​​corresponding to the difference is determined as the dry mass.

[0100] The following uses needle-shaped tea as an example to illustrate the implementation of the embodiments of this application in detail.

[0101] Figure 1 An exemplary flow diagram of the image recognition-based moisture detection method for needle-shaped tea leaves in an embodiment of this application is shown. See also Figure 1 As shown, the image recognition-based moisture detection method for needle-shaped tea leaves in this application embodiment may include:

[0102] Step S101: Obtain the humidity detection result of the image of the needle-shaped tea to be detected. The humidity detection result includes at least two detection parameters.

[0103] The indicators corresponding to the detection parameters can characterize a moisture feature during the needle-shaped tea processing. This moisture feature can include, but is not limited to, moisture content characteristics, shape parameter characteristics, morphological characteristics, or any other characteristics related to the state of the needle-shaped tea during processing. In practical applications, the indicator can be pre-configured, selected in real-time, or chosen in response to the needle-shaped tea operation, taking into account the specific application scenario, the condition of the needle-shaped tea, the requirements for needle-shaped tea, the accuracy requirements, or other factors. For example, in complex growing environments and harsh weather conditions, more indicators can be selected to accurately detect the moisture morphology of the needle-shaped tea, ensuring accurate needle-shaped tea detection. Furthermore, for different types of needle-shaped tea, due to the different requirements during processing, the indicators need to be adjusted according to the different shape parameters of the needle-shaped tea.

[0104] The detection parameters can be data obtained after preprocessing such as real-time cleaning and normalization from real-time acquired indicator detection commands. In some examples, if the probability value of the processing range is [0,1], the detection parameter can be a normalized value obtained based on the real-time acquired indicator detection commands, with the normalized value taking values ​​between [0,1]. This not only facilitates efficient determination of whether the detection parameter falls within the processing range by the device (e.g., the computing device mentioned below), but also reduces the number of pixels and computational complexity. In other examples, the detection parameters can also be integer values ​​obtained based on real-time acquired indicator detection commands. This not only facilitates efficient determination of whether the detection parameter falls within the corresponding processing range, but also reduces the requirements for hardware performance such as processors, further reducing hardware costs.

[0105] The indicator detection commands can be acquired in real time by sensors. The indicator detection commands can be, but are not limited to, numerical detection commands, simulated detection commands, or other similar detection commands. These detection commands themselves have a small number of pixels and are not affected by the acquisition environment. The data processing and transmission have low requirements for computing power and bandwidth, which can not only reduce computational complexity, but also reduce the application cost of the method in this embodiment.

[0106] In at least some embodiments, the indicators may include, but are not limited to, moisture content indicators, shape parameter indicators, etc. In this embodiment, the moisture characteristics of needle-shaped tea can be comprehensively and accurately perceived through multiple indicators, and humidity detection results that can reflect the comprehensive moisture characteristics of needle-shaped tea can be obtained, thereby efficiently and accurately determining the moisture morphology of needle-shaped tea.

[0107] Moisture content indicators can indicate the status of processing-related biological monitoring requirements. In some implementations, moisture content indicators may include, but are not limited to, pre-drying mass, mass, temperature, oxygen concentration, or fermentation rate.

[0108] Moisture content detection commands can be acquired by a biosensor. Here, the biosensor can provide these commands to the device (e.g., a computing device hereinafter) via various wireless processing methods, or via a wired connection to the device in the form of electrical detection commands.

[0109] Biological detection sensors can include, but are not limited to, respiration rate sensors, mass sensors, fermentation rate sensors, oxygen concentration sensors, and temperature sensors. Specifically, respiration rate sensors can be used to collect respiration rate data in real time during the tea-making process; mass sensors can be used to collect mass data in real time; temperature sensors can be used to collect temperature data in real time; fermentation rate sensors can be used to collect fermentation rate data in real time; and oxygen concentration sensors can be used to collect oxygen concentration data in real time. Typically, either fermentation rate or oxygen concentration can be selected.

[0110] Shape parameters can indicate the status of tea drinking actions for needle-shaped tea leaves. In some implementations, shape parameters may include, but are not limited to, the curling rate of needle-shaped tea leaves, the curling degree of needle-shaped tea leaves, the area of ​​needle-shaped tea leaves, and the thickness of needle-shaped tea leaves. The detection instructions for shape parameters may include, but are not limited to, instructions for detecting the curling rate of needle-shaped tea leaves and instructions for detecting the area of ​​needle-shaped tea leaves.

[0111] The shape parameter index detection commands can be acquired by a shape parameter sensor. The shape parameter sensor can provide its real-time acquired detection commands to the device (e.g., the computing device below) through various wireless processing methods, or it can provide its real-time acquired detection commands to the device (e.g., the computing device below) in the form of electrical detection commands through a wired connection with the device.

[0112] It is understood that the types of indicators in the embodiments of this application are not limited to the shape parameters and moisture content mentioned above, but may also include other types that can characterize the moisture status of needle-shaped tea when it is brewed, such as indicators related to voice and pupil. For example, the indicators of needle-shaped tea may also include language indicators. The detection parameters of voice may include the frequency value, amplitude value, etc. of the language indicator, and the language indicator detection instructions can be collected by voice sensors such as microphone arrays.

[0113] Humidity detection results are a collection of various detection parameters at the same time. These results have at least two dimensions, each corresponding one-to-one with an indicator, reflecting the overall moisture characteristics of needle-shaped tea at a given moment. For example, taking six pre-selected indicators—respiration rate, mass, oxygen concentration (or fermentation rate), needle-shaped tea leaf curling rate, and needle-shaped tea leaf area—as an example, the corresponding humidity detection results can have six dimensions, each representing one of the six indicators. The humidity detection result at time i can be represented as {ai,bi,ci,di,ei}, where ai represents the respiration rate at time i, bi represents the mass at time i, ci represents the oxygen concentration or fermentation rate at time i, di represents the needle-shaped tea leaf curling rate at time i, and ei represents the needle-shaped tea leaf area at time i. It should be noted that "the same time" here includes not only strict time synchronization but also cases where the acquisition time interval is less than a predetermined threshold. This predetermined threshold can be determined by the performance parameters of each sensor. In practical applications, this threshold can be an empirical value or a set value.

[0114] Step S102: When the number of images to be detected for the detection parameter that falls within the range of the fixation process is greater than a first predetermined value among the at least two detection parameters of the humidity detection result, the humidity morphology corresponding to the fixation process range value is determined as the humidity morphology of needle-shaped tea.

[0115] Here, the term "blanching treatment range value" refers generally to the range value of a single type of treatment. This range value can be used to indicate a single humidity profile and can include treatment range values ​​that correspond one-to-one with each indicator. Detailed technical specifications are described below and will not be repeated here.

[0116] Humidity patterns can indicate the moisture state associated with drinking needle-shaped tea. In this embodiment, humidity patterns may include two or more types. Thus, by classifying humidity patterns, needle-shaped tea can intuitively understand whether its state is suitable for drinking, while reducing the reliance on supervised machine learning models in the state detection process, effectively reducing computational complexity while improving accuracy.

[0117] In at least some embodiments, the humidity profile can include at least two types. Correspondingly, each needle-shaped tea can have at least two processing range values ​​corresponding to each humidity profile, and each withering processing range value corresponds to one of the humidity profiles. In this embodiment, the various processing range values ​​of the needle-shaped tea are directly related to each humidity profile. Its humidity profile can be directly determined through at least two processing range values ​​without relying on intermediate features such as color or brightness type. This reduces computational complexity, decreases false or missed identifications, and improves the accuracy of humidity profile detection.

[0118] In some implementations, the humidity profile can include a pre-treatment humidity profile, a slightly dry humidity profile, a moderately dry humidity profile, and a heavily dry humidity profile. Each needle-shaped tea can have four processing range values ​​corresponding to these four humidity profiles, i.e., four fixation processing range values. These four fixation processing range values ​​correspond to the pre-treatment humidity profile, the slightly dry humidity profile, the moderately dry humidity profile, and the heavily dry humidity profile, respectively. In this example, the humidity profile of the needle-shaped tea can be directly determined through multiple processing range values, without relying on intermediate features such as color or brightness type. This reduces computational complexity and further reduces false or missed identifications.

[0119] In some implementations, the humidity profile can include both a pending humidity profile and a dry humidity profile. The processing range values ​​for needle-shaped tea can include one type of processing range value corresponding to the pending humidity profile and another type of processing range value corresponding to the dry humidity profile; that is, two processing range values ​​for fixing (or killing the green) that correspond to the pending humidity profile and the dry humidity profile, respectively. In this example, the humidity profile of the needle-shaped tea can be directly determined as pending or dry using only these two types of processing range values, without relying on intermediate features such as color or brightness type. This reduces computational complexity and the possibility of misidentification or missed identification.

[0120] In at least some embodiments, each type of processing range value may include a processing range value that corresponds one-to-one with the index; that is, the range value of the fixation processing may include a processing range value that corresponds one-to-one with the index. Therefore, the humidity morphology of needle-shaped tea can be determined by comprehensively considering various moisture characteristics during processing, which can further improve the accuracy of humidity morphology detection for needle-shaped tea.

[0121] Taking six pre-selected indicators as an example, these six indicators include respiration rate, mass, oxygen concentration (or, fermentation rate), needle-shaped tea leaf curling rate, needle-shaped tea leaf area, and needle-shaped tea leaf curling degree. Assuming the humidity profile includes both the untreated humidity profile and the dry humidity profile, the corresponding treatment range values ​​are divided into two categories: one category corresponds to the untreated humidity profile, and the other corresponds to the dry humidity profile. In this example, each category of treatment range values ​​includes six treatment range values, which correspond to respiration rate, mass, oxygen concentration (…), and… Alternatively, the range of values ​​for each treatment category can be expressed as P = {(La,Ua),(Lb,Ub),(Lc,Uc),(Ld,Ud),(Le,Ue),(Lf,Uf)}, where L represents the lower limit, U represents the upper limit, subscript a represents the respiration rate, subscript b represents the mass, subscript c represents the oxygen concentration or fermentation rate, subscript d represents the needle-shaped tea leaf curling rate, subscript e represents the needle-shaped tea leaf area, and subscript f represents the needle-shaped tea leaf curling degree.

[0122] In step S102, if the value of the detected parameter in the humidity detection result belongs to the processing range value corresponding to the corresponding indicator in the fixation processing range value, then the detected parameter can be considered to fall within the fixation processing range value. If the number of images to be detected for a detected parameter that falls within the fixation processing range value in a humidity detection result is greater than a first predetermined value, then the humidity detection result can be considered to fall within the fixation processing range value of needle-shaped tea. In this embodiment, the humidity state of needle-shaped tea can be determined by judging whether a specific value belongs to the corresponding value range. The algorithm is simple and easy to implement, has low hardware cost, and is highly efficient with small error.

[0123] Analysis revealed that even when the moisture content of needle-shaped tea is to be processed, some of its detection parameters may still fall within the range of values ​​for the fixation treatment corresponding to the moisture content of the dried tea. Similarly, when the moisture content of needle-shaped tea is dried, some of its detection parameters may also fall within the range of values ​​for the fixation treatment corresponding to the moisture content of the tea to be processed. Therefore, the method in this embodiment determines the moisture content of needle-shaped tea by setting a first predetermined value in conjunction with the processing range value, which is more consistent with reality and yields more accurate results.

[0124] In practical applications, the first predetermined value can be determined in various ways, such as pre-configuration, real-time selection, or selection in response to needle-shaped tea operations, depending on the specific application scenario, the condition of the needle-shaped tea, the accuracy requirements, the requirements for accurate needle-shaped tea detection, the growth environment, weather, surrounding environment, and / or other factors. In some embodiments, the first predetermined value can be a default value, which can be an empirical value obtained through statistical analysis. In some embodiments, the first predetermined value can be a fixed value or a function value of a user-defined function with the above-mentioned factors and the current total number of indicators as variables. For example, the first predetermined value can be set to one-half, two-thirds, or three-quarters of the total number of indicators. Taking the pre-selection of 6 indicators as an example, the first predetermined value can be set to 3 or 4.

[0125] In at least some embodiments, the image recognition-based method for detecting the humidity of needle-shaped tea may further include: step S103, where, among at least two detection parameters in the humidity detection result, the number of images to be detected for a detection parameter falling within the range of the withering process is less than or equal to a first predetermined value, the humidity profile of the needle-shaped tea is determined based on the difference between the humidity detection result and a first reference humidity parameter value for needle-shaped tea. Thus, the difference between the humidity detection result and the first reference humidity parameter value for needle-shaped tea corresponding to different humidity profiles can be used to assist in the detection of the humidity profile, further achieving real-time, efficient, and accurate detection of the humidity profile of needle-shaped tea with lower computational complexity.

[0126] The first needle-shaped tea reference humidity parameter value generally refers to a needle-shaped tea reference humidity parameter value corresponding to a single humidity morphology. In other words, a single processing range value can correspond to a needle-shaped tea reference humidity parameter value, or a withering processing range value corresponds to a first needle-shaped tea reference humidity parameter value. In some implementations, similar to the processing range value, each first needle-shaped tea reference humidity parameter value can correspond to a humidity morphology. The first needle-shaped tea reference humidity parameter value has at least two values ​​that correspond one-to-one with at least two indicators. That is, the first needle-shaped tea reference humidity parameter value can be represented by multi-dimensional data, where the dimension of the multi-dimensional data is equal to the number of indicator images to be detected. Specifically, each dimension of the first needle-shaped tea reference humidity parameter value corresponds to an indicator, and the value of each dimension can be used as the mean value of the corresponding indicator. Therefore, by using the values ​​of each dimension of the needle-shaped tea reference humidity parameter value and the detection parameters in the humidity detection results of the needle-shaped tea images to be detected, the difference between the humidity detection result and the needle-shaped tea reference humidity parameter value can be accurately estimated. In this implementation, the humidity pattern of needle-shaped tea is determined by the difference between multi-dimensional data. The humidity pattern of needle-shaped tea can be determined by the comprehensive characteristics of humidity during processing, which improves the accuracy of humidity pattern detection while reducing computational complexity.

[0127] In some embodiments, similar to the range value for the fixation process, the first needle-shaped tea reference humidity parameter value may include two or more, each corresponding to a humidity mode. Taking the humidity mode including the humidity mode to be treated and the dry humidity mode as an example, the needle-shaped tea reference humidity parameter value may include a needle-shaped tea reference humidity parameter value corresponding to the humidity mode to be treated and a needle-shaped tea reference humidity parameter value corresponding to the dry humidity mode. That is, one or two first needle-shaped tea reference humidity parameter values ​​can be set, and these two first needle-shaped tea reference humidity parameter values ​​correspond to the humidity mode to be treated and the dry humidity mode, respectively. In other words, the first needle-shaped tea reference humidity parameter value can correspond to either the humidity mode to be treated or the dry humidity mode. Therefore, the humidity mode of the needle-shaped tea can be determined by the difference between the humidity detection result and the first needle-shaped tea reference humidity parameter value, without relying on intermediate features such as color and brightness type, reducing computational complexity and further reducing the possibility of misidentification or omission.

[0128] In some embodiments, in step S103, among two first needle-shaped tea reference humidity parameter values, the humidity pattern corresponding to the first needle-shaped tea reference humidity parameter value that is closer to the humidity detection result is determined as the humidity pattern of the needle-shaped tea. Alternatively, among three or more first needle-shaped tea reference humidity parameter values, the humidity pattern corresponding to the first needle-shaped tea reference humidity parameter value that is closest to the humidity detection result is determined as the humidity pattern of the needle-shaped tea. In this embodiment, the needle-shaped tea reference humidity parameter value closest to the humidity detection result is obtained by the difference between the humidity detection result and the needle-shaped tea reference humidity parameter value, and then the humidity pattern corresponding to the needle-shaped tea reference humidity parameter value closest to the humidity detection result is used as the humidity pattern of the needle-shaped tea, which has low computational complexity and small error.

[0129] In some embodiments, if multiple fixation processing range values ​​exist, in step S103, when the number of images to be detected falling into each fixation processing range value among at least two detection parameters of the humidity detection result is less than or equal to a first predetermined value, the humidity profile of the needle-shaped tea is determined based on the difference between the humidity detection result and the first needle-shaped tea reference humidity parameter value. Thus, by preferentially using the fixation processing range value and the first needle-shaped tea reference humidity parameter value as auxiliary means, the humidity profile can be determined, which is beneficial for improving efficiency and accuracy.

[0130] In some implementations, the difference between the humidity detection result and the reference humidity parameter value for the first needle-shaped tea can be, but is not limited to, a Euclidean difference, a simple difference, etc. In some examples, the difference between the humidity detection result and the reference humidity parameter value for the first needle-shaped tea can be a simple difference, which can reduce the computational cost and eliminate the error caused by approximation operations such as square root extraction.

[0131] In some implementations, step S103, which determines the humidity profile of the needle-shaped tea based on the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea, may include: step a1, where the absolute value of the difference between the first difference and the second difference is greater than the average of the first and second differences, the humidity profile is determined to be the humidity profile to be processed; step a2, where the absolute value of the difference between the first difference and the second difference is less than or equal to the average of the first and second differences, the humidity profile is determined to be a dry humidity profile. The first difference is the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea corresponding to the humidity profile to be processed, and the second difference is the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea corresponding to the dry humidity profile. Therefore, the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea can be determined using a simple algorithm that calculates the average and the difference. This reduces computational complexity and error, further lowering hardware costs while improving the accuracy of humidity profile detection.

[0132] In some implementations, where the accuracy requirement for humidity pattern detection is relatively low, the humidity pattern corresponding to the first needle-shaped tea reference humidity parameter value, after estimating the difference between the humidity detection result and the first needle-shaped tea reference humidity parameter value with the smallest difference, can be determined as the humidity pattern of the needle-shaped tea. This implementation method has lower computational complexity and relatively lower hardware overhead.

[0133] In at least some embodiments, the processing range value and the reference humidity parameter value for the first needle-shaped tea can be obtained synchronously using various applicable algorithms. For example, clustering or other unsupervised learning algorithms can be used to determine the processing range value and the reference humidity parameter value for the needle-shaped tea. As another example, the processing range value and the reference humidity parameter value for the first needle-shaped tea can be determined using data analysis and statistical algorithms. The specific algorithms used to determine the processing range value and the reference humidity parameter value for the first needle-shaped tea are not limited in the embodiments of this application.

[0134] In some implementations, the range of values ​​for the initial processing (kill-green) and the corresponding reference humidity parameter value for the first needle-shaped tea can be determined using an unsupervised learning algorithm. This unsupervised learning algorithm can be, but is not limited to, clustering, density estimation, and autoencoders. Compared to supervised classification algorithms such as SVM in related technologies, using unsupervised clustering algorithms offers the feasibility of setting values ​​based on the specific type of needle-shaped tea. That is, the processing range and reference humidity parameter values ​​can be determined according to the specific type of needle-shaped tea, thereby improving the accuracy of personalized humidity morphology detection and solving the problem of applying the same standard to all needle-shaped teas in related technologies.

[0135] In some implementations, the range value of the fixation process and the reference humidity parameter value of the first needle-shaped tea can be obtained based on the humidity detection result library of the needle-shaped tea itself. The humidity detection result library contains the humidity detection results of the needle-shaped tea images to be tested obtained in advance. In one implementation, the range value of the fixation process and the reference humidity parameter value of the first needle-shaped tea can be determined by clustering the existing humidity detection results in the needle-shaped tea humidity detection result library. Specifically, the range value of the fixation process can be obtained as follows: Step b1, clustering is performed based on the humidity detection result library of the needle-shaped tea images to be tested to form a needle-shaped tea image set corresponding to the humidity pattern. The humidity detection result library of the needle-shaped tea images to be tested contains the humidity detection results of the needle-shaped tea images to be tested obtained in advance; Step b2, data configuration filtering is performed on the needle-shaped tea image set to determine whether the data configuration of the needle-shaped tea image set belongs to a balanced configuration or an unbalanced configuration; Step b3, the range value of the fixation process corresponding to the humidity pattern is determined according to the data configuration of the needle-shaped tea image set. In this implementation, the reference humidity parameter value of the needle-shaped tea in the image set corresponding to the humidity morphology is the first reference humidity parameter value of the needle-shaped tea corresponding to the range value of the fixation process, which can be determined in the clustering in step b1. Thus, by clustering based on the humidity detection result library of the needle-shaped tea itself to obtain the range value of the fixation process and the first reference humidity parameter value of the needle-shaped tea, not only are the set values ​​of the fixation process range value and the first reference humidity parameter value of the needle-shaped tea adjusted according to the specific type of needle-shaped tea, thereby achieving personalization or customization of humidity morphology detection and improving the accuracy of humidity morphology detection, but also reducing missed or false identifications.

[0136] In some implementations, data configuration filtering of the needle-shaped tea image set can be achieved through the following steps: Step c1, the humidity detection results in the needle-shaped tea image set are divided into pixel grids to obtain multiple pixel grids, each pixel grid having the same number of pixels; Step c2, the pixel grid with the largest number of pixels among the multiple pixel grids is taken as the mean range of the needle-shaped tea image set. When the number of pixels in the first range of the needle-shaped tea image set is equal to the number of pixels in the second range, the data configuration of the needle-shaped tea image set is a balanced configuration; when the number of pixels in the first range of the needle-shaped tea image set is not equal to the number of pixels in the second range, the data configuration of the needle-shaped tea image set is an unbalanced configuration; wherein, the first range and the second range are obtained by dividing the data range of the needle-shaped tea image set according to the mean range. In this implementation, the range of pixel grid division and data counting, as well as the comparison of the number of data on the left and right halves of the mean range, are used to determine whether the needle-shaped tea image set is a balanced or unbalanced configuration. This approach combines low complexity with accurate data configuration filtering, which can further reduce hardware costs and improve the accuracy of needle-shaped tea moisture morphology detection.

[0137] The drinking methods for the same needle-shaped tea may vary significantly at different times (e.g., at different age groups). To address similar or other situations, in at least some embodiments, the range value of the fixation treatment and the reference humidity parameter value of the first needle-shaped tea can be adjusted in real time as needed. This allows the fixation treatment range value and the reference humidity parameter value of the first needle-shaped tea to dynamically change with the changes in the needle-shaped tea itself, thereby improving the accuracy of humidity morphology detection at different times for the same needle-shaped tea, i.e., achieving time-varying humidity morphology detection.

[0138] In some implementations, the process of adjusting the range value of the fixation treatment and the reference humidity parameter value of the first needle-shaped tea may include: updating the humidity detection result library of the needle-shaped tea images to be detected based on the newly acquired humidity detection results, and redetermining the range value of the fixation treatment using the unclassified humidity detection results in the updated humidity detection result library. Here, unclassified humidity detection results refer to humidity detection results in the needle-shaped tea image set that are not classified into each humidity morphology, that is, humidity detection results that do not fall within the fixation treatment range value of the needle-shaped tea. Here, the algorithm used to redetermine the range value of the fixation treatment and the reference humidity parameter value of the first needle-shaped tea is the same as the related algorithm mentioned above, and will not be repeated. In this embodiment, the adjustment of the range value of the fixation treatment and the reference humidity parameter value of the first needle-shaped tea is achieved using unclassified humidity detection results, without having to repeatedly process all humidity detection results of the needle-shaped tea. The update of the range value of the fixation treatment and the reference humidity parameter value of the first needle-shaped tea can be completed with less computation and lower computational complexity, realizing the dynamic adjustment of the range value of the fixation treatment and the reference humidity parameter value of the first needle-shaped tea, that is, the humidity morphology detection can be made different over time with lower hardware cost.

[0139] In some implementations, the humidity detection results of the needle-shaped tea image to be tested can be stored in a humidity detection result database of the needle-shaped tea image to be tested in the form of a time data sequence. The humidity detection result database can be read and written by verifying preset information, which can include any information related to identifying a specific needle-shaped tea. This application does not limit the type of preset information. In addition, the processing range value and the reference humidity parameter value of the needle-shaped tea can also be stored in the humidity detection result database of the needle-shaped tea image to be tested. In this way, the processing range value and the reference humidity parameter value of the first needle-shaped tea can be queried in real time through the preset information of the needle-shaped tea, and the humidity detection results of the needle-shaped tea image to be tested can be updated. This realizes the detection of the humidity state of the needle-shaped tea anytime and anywhere, without being limited by processing conditions, region, equipment, needle-shaped tea, or other usage conditions.

[0140] In this embodiment of the application, the humidity detection result database can be updated in various applicable ways.

[0141] In some implementations, each new humidity detection result is directly added to the humidity detection result library of the needle-shaped tea images to be inspected. This exemplary implementation facilitates the rapid formation of the humidity detection result library of needle-shaped tea images to be inspected. This exemplary implementation is suitable when the humidity detection result inclusion rate of the needle-shaped tea images to be inspected is low, or when the humidity detection result library of needle-shaped tea images to be inspected is initially created.

[0142] In some implementations, when the humidity detection result inclusion rate of the needle-shaped tea image to be tested is greater than or equal to a third predetermined value, humidity detection results with a number of images falling within the range of the withering process that are less than or equal to a first predetermined value are added to the humidity detection result database of the needle-shaped tea image to be tested. Thus, when the range of the withering process for needle-shaped tea and the first reference humidity parameter value for needle-shaped tea are relatively stable, an added mechanism for judging the entry of humidity detection results into the database can reduce invalid or duplicate data, avoid unnecessary redundant calculations, and improve the data integrity of the needle-shaped tea humidity detection result database. This allows for the continuous improvement of needle-shaped tea images while reducing redundant data in the humidity detection result database, lowering the amount of data computation, and reducing computational complexity.

[0143] In the above implementation, the humidity detection result inclusion rate refers to the proportion of humidity detection results that fall within the range of the blanching process out of all humidity detection results. A humidity detection result falling within the range of the blanching process refers to a humidity detection result where the number of images containing the detection parameter within the blanching process range is greater than a first predetermined value. In practical applications, the humidity detection result inclusion rate of the images containing the humidity detection results that fall within the range of the blanching process can be calculated in real time.

[0144] In practical applications, the specific value of the third predetermined value can be pre-configured, dynamically determined, or selected in response to needle-shaped tea operations, based on one or more factors such as the needs of the specific application scenario, the requirements for detection accuracy, the relevant regulations for accurate tea drinking, the demand for needle-shaped tea, and the specific situation of needle-shaped tea. In some embodiments, the third predetermined value can be an empirical value, the function value of a predefined function with the above factors as variables, etc. The third predetermined value can be a fixed value, a range value, or a variable value. In some examples, the third predetermined value can be a percentage between 0 and 1, such as 90%, 80%, 70%, etc. Taking a humidity detection result inclusion rate of 90% as an example, if 9 out of every 10 humidity detection results fall within the existing processing range value, the entry strategy of the humidity detection result database can be changed to "only humidity detection results that do not fall within the processing range value can enter the humidity detection result database".

[0145] In at least some embodiments, the calculation and updating of the withering range value and the first needle-shaped tea reference humidity parameter value can be performed locally at the needle-shaped tea detection terminal or via a cloud server. In some implementations, the calculation and updating of the withering range value and the first needle-shaped tea reference humidity parameter value can be obtained from the cloud server and then provided to the needle-shaped tea detection terminal. In this implementation, the calculation of the withering range value and the first needle-shaped tea reference humidity parameter value is performed on the cloud server, while the detection of humidity characteristics is performed locally at the needle-shaped tea detection terminal.

[0146] In at least some embodiments, the method of this application may further include a step of adjusting the humidity profile. After detecting that the needle-shaped tea is in a dry humidity profile, the humidity profile of the needle-shaped tea can be adjusted by single or multiple detection commands, thereby improving the dry humidity profile of the needle-shaped tea in a timely and effective manner.

[0147] In some implementations, the method described in this application embodiment may further include: step S104, when the duration of the dry humidity state reaches a second predetermined value, adjusting the humidity state using one or more adjustment mechanisms including the following: initial kneading, re-kneading, second fermentation, and hot air dehydration. In this implementation, triggering the adjustment mechanism after a certain period of time not only enables timely and effective intervention in the dry humidity state but also avoids accidental triggering of the adjustment mechanism, thus improving the needle-shaped tea experience.

[0148] In practical applications, the specific value of the second predetermined value can be pre-configured, dynamically determined, or selected in response to needle-shaped tea operations, based on one or more factors such as the needs of the specific application scenario, the requirements for detection accuracy, relevant regulations for accurate tea drinking, the needs of needle-shaped tea, and the condition of the needle-shaped tea. In some embodiments, the second predetermined value can be an empirical value, the function value of a predefined function with the above-mentioned factors as variables, etc. The second predetermined value can be a fixed value, a range value, or a variable value. For example, the second predetermined value can be 5 seconds, that is, if the duration of the needle-shaped tea's humidity state being dry is greater than or equal to 5 seconds, the adjustment mechanism for the humidity state of the needle-shaped tea is triggered. It should be noted that "5 seconds" here is only an example; in actual applications, the second predetermined value can be 10 seconds or more, more than one minute, or even longer, and can be freely set.

[0149] In some implementations, the adjustment mechanism can be determined based on the drying priority of the needle-shaped tea. This drying priority can be determined by the humidity detection results of the needle-shaped tea's image and the range of values ​​corresponding to its humidity morphology during the fixation process. For different drying priorities, the drying priority of the needle-shaped tea can be redefined after different time intervals. That is, the first drying priority of the needle-shaped tea is redefined after the first time interval, with the first time interval corresponding to the first drying priority. In this way, by using an appropriate method to adjust the state for different degrees of dryness, the effect of humidity morphology intervention can be significantly improved.

[0150] Here, "first drying priority" generally refers to a single drying priority, and "first time interval" generally refers to a time interval that corresponds to or is associated with a single drying priority.

[0151] In some examples, the range of values ​​for the fixation treatment of needle-shaped tea can be divided into multiple sub-ranges, each corresponding to a different drying priority, and a different adjustment mechanism can be configured for each drying priority. The drying priority corresponding to the sub-range of the fixation treatment range value where the humidity detection result of the needle-shaped tea's image to be detected falls is the drying priority of the needle-shaped tea. The adjustment mechanism corresponding to this drying priority can be used to adjust the humidity profile of the needle-shaped tea.

[0152] In some implementations, the mode of regulation can be randomly selected. That is, a single-mode or multi-mode regulation mechanism can be used, and the specific mode used during regulation can be randomly selected, which can effectively improve the resilience of the regulation mechanism.

[0153] The image recognition-based method for detecting the humidity of needle-shaped tea in this application obtains the humidity detection result of the image of the needle-shaped tea to be detected. By judging the number of images of the detection parameter that fall within the processing range value of at least two detection parameters of the humidity detection result, the humidity state of the needle-shaped tea can be directly determined. It has low computational complexity and the algorithm is easy to implement. It can accurately, efficiently and in real time detect the humidity state of needle-shaped tea with low computational complexity, while improving the needle-shaped tea experience and reducing application costs.

[0154] The image recognition-based humidity detection method for needle-shaped tea in this application is highly reusable and can be used to detect various states of needle-shaped tea during tea drinking. The same method and apparatus can be reused to detect brightness state, fatigue state, safety state, and tea drinking loop state, and can also assist in the accurate detection of humidity and the safe detection of needle-shaped tea.

[0155] Figure 2 An exemplary structure of the image recognition-based needle-shaped tea moisture detection system provided in this application embodiment is shown. See also Figure 2 An exemplary image recognition-based needle-shaped tea moisture detection system in the illustrated embodiment may include:

[0156] Acquisition device 21 is used to acquire the humidity detection result of the image to be detected of the needle-shaped tea, and the humidity detection result includes at least two detection parameters;

[0157] The determining device 22 is used to determine the humidity morphology corresponding to the fixation processing range value as the humidity morphology of needle-shaped tea when the number of images to be detected for the detection parameter falling within the fixation processing range value is greater than a first predetermined value among at least two detection parameters.

[0158] The image recognition-based needle-shaped tea humidity detection system of this application acquires the humidity detection results of the needle-shaped tea image to be detected. By judging the number of detection parameters of the image to be detected that fall within the processing range value of at least two detection parameters of the humidity detection results, the humidity state of the needle-shaped tea can be directly determined. It has low computational load and the algorithm is easy to implement. It can accurately, efficiently and in real time detect the humidity state of needle-shaped tea with low computational complexity, while improving the needle-shaped tea experience and reducing application costs.

[0159] In some implementations, at least two detection parameters are used to indicate moisture content and / or shape parameters. In other words, the indicators corresponding to the detection parameters include moisture content indicators and / or shape parameter indicators. Therefore, the moisture status of needle-shaped tea can be comprehensively and accurately perceived through multiple indicators, thereby efficiently and accurately determining the moisture content of the tea.

[0160] In some implementations, the humidity profile includes a pre-processing humidity profile and a dry humidity profile, with the range value of the fixation process corresponding to either the pre-processing humidity profile or the dry humidity profile. Therefore, by using the two types of processing range values ​​for needle-shaped tea, it is possible to directly determine whether the humidity profile of the needle-shaped tea is pre-processing or dry, without relying on intermediate features such as color or brightness. This reduces computational complexity and minimizes misidentification or missed identification.

[0161] In some implementations, the processing range value for each type of processing, i.e., the processing range value for each fixation process, can include processing range values ​​that correspond one-to-one with the indicators corresponding to the detection parameters. Therefore, it is possible to comprehensively determine the moisture content of needle-shaped tea by considering various moisture characteristics during processing.

[0162] In some implementations, the image recognition-based needle-shaped tea humidity detection system of this application embodiment may further include a computing device 23. The range value for the fixation processing is determined by the computing device 23, which can be configured to obtain the fixation processing range value through the following steps: clustering based on a humidity detection result library of needle-shaped tea images to form a needle-shaped tea image set corresponding to the humidity morphology; the humidity detection result library of needle-shaped tea images to be detected contains pre-obtained humidity detection results of the needle-shaped tea images to be detected; filtering the data configuration of the needle-shaped tea image set to determine whether the data configuration of the needle-shaped tea image set is a balanced configuration or an unbalanced configuration; and determining the fixation processing range value corresponding to the humidity morphology based on the data configuration of the needle-shaped tea image set. Thus, the processing range value and the reference humidity parameter value of the needle-shaped tea are set according to the specific type of needle-shaped tea, while reducing missed or false identifications.

[0163] In some implementations, the computing device 23 may also be configured to perform data configuration filtering on the needle-shaped tea image set through the following steps: dividing the humidity detection results in the needle-shaped tea image set into pixel grids to obtain multiple pixel grids, each pixel grid having the same number of pixels; using the pixel grid with the largest number of pixels among the multiple pixel grids as the mean range of the needle-shaped tea image set; when the number of pixels in the first range of the needle-shaped tea image set is equal to the number of pixels in the second range of the needle-shaped tea image set, the data configuration of the needle-shaped tea image set is a balanced configuration; when the number of pixels in the first range of the needle-shaped tea image set is not equal to the number of pixels in the second range of the needle-shaped tea image set, the data configuration of the needle-shaped tea image set is an unbalanced configuration; wherein, the first range and the second range are obtained by dividing the data range of the needle-shaped tea image set according to the mean range. Therefore, the image recognition-based needle-shaped tea moisture detection system of this application determines whether the needle-shaped tea image set is balanced or unbalanced by dividing the range of pixel grids and counting data, and comparing the number of data on the left and right halves of the mean range. It has the advantages of low complexity and accurate data configuration screening, which can further reduce hardware costs and improve the accuracy of needle-shaped tea moisture morphology detection.

[0164] In some implementations, the determining device 22 can also be used to determine the humidity profile of the needle-shaped tea based on the difference between the humidity detection result and the first needle-shaped tea reference humidity parameter value when the number of images to be detected for the detection parameter falling within the range of the withering process is less than or equal to a first predetermined value among at least two detection parameters of the humidity detection result.

[0165] In some embodiments, the first needle-shaped tea reference humidity parameter value may include two or more, and each first needle-shaped tea reference humidity parameter value corresponds to a humidity pattern. The determining device 22 is specifically used to determine the humidity pattern of the needle-shaped tea by choosing the first needle-shaped tea reference humidity parameter value that is closer to the humidity detection result from two first needle-shaped tea reference humidity parameter values; or, from three or more first needle-shaped tea reference humidity parameter values, to determine the humidity pattern of the needle-shaped tea by choosing the first needle-shaped tea reference humidity parameter value that is closest to the humidity detection result. Thus, the difference between the humidity detection result of the needle-shaped tea's image to be detected and its needle-shaped tea reference humidity parameter value can assist in the detection of the needle-shaped tea's humidity pattern, achieving real-time, efficient, and accurate detection of the needle-shaped tea's humidity pattern with low computational complexity.

[0166] In some implementations, the determining device 22 can be specifically used to: determine the humidity pattern as the humidity pattern to be processed when the difference between the first difference and the second difference is greater than the average of the first difference and the second difference; and determine the humidity pattern as a dry humidity pattern when the difference between the first difference and the second difference is less than or equal to the average of the first difference and the second difference. The first difference is the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea corresponding to the humidity pattern to be processed, and the second difference is the difference between the humidity detection result and the reference humidity parameter value of the first needle-shaped tea corresponding to the dry humidity pattern. In this way, the humidity pattern of the needle-shaped tea can be determined based on the difference with low computational complexity.

[0167] In some implementations, the image recognition-based needle-shaped tea humidity detection system of this application embodiment may further include: a database update device 24, used to add humidity detection results whose number of detection images falling within the range of the withering process is less than or equal to a first predetermined value to the humidity detection result database of the needle-shaped tea images when the humidity detection result inclusion rate of the images to be detected is greater than or equal to a third predetermined value. This significantly reduces redundant data entering the humidity detection result database, lowering the amount of data computation and computational complexity.

[0168] In some implementations, the image recognition-based needle-shaped tea humidity detection system of this application embodiment may further include: an adjustment device 25, used to adjust the humidity state by employing one or more adjustment mechanisms including the following methods when the duration of the humidity state being dry is greater than or equal to a second predetermined value: initial rolling, re-rolling, second withering, and hot air dehydration. This allows for timely and effective intervention in the dry humidity state of the needle-shaped tea, while avoiding false triggering of the adjustment mechanism.

[0169] In some implementations, the adjustment mechanism can be determined based on the drying priority of the needle-shaped tea. This drying priority can be determined by the humidity detection results of the image of the needle-shaped tea to be tested and the range of values ​​for the fixation treatment corresponding to the humidity morphology. In this way, by using appropriate methods to adjust the state for different degrees of dryness, the effectiveness of the humidity morphology intervention can be significantly improved.

[0170] In some implementations, the adjusting device 25 can also be used to redetermine the first drying priority of the needle-shaped tea after the first time interval, with the first time interval corresponding to the first drying priority. In this way, re-evaluating at different time intervals for different degrees of dryness is more in line with the actual situation of improving the moisture morphology of needle-shaped tea, and can significantly improve the effect of moisture morphology intervention.

[0171] In some implementations, the mode of regulation can be randomly selected by the regulation device 25. This can effectively improve the resistance of the regulation mechanism.

[0172] The following provides a detailed description of some exemplary embodiments of this application. The various embodiments described below can be arbitrarily combined with each other, and the various exemplary implementations of each embodiment can also be arbitrarily combined as needed. It should be noted that the various embodiments described below and their arbitrary combinations are merely examples of the methods or apparatus described in the embodiments of this application, and are not intended to limit the embodiments of this application.

[0173] Figure 3 An exemplary process for humidity morphology detection according to a second embodiment of this application is illustrated. This exemplary process may include the following steps:

[0174] Step S301: During the processing of needle-shaped tea, the needle-shaped tea detection terminal device installed on the needle-shaped tea obtains the preset information of the needle-shaped tea and the numerical detection instructions of its various moisture content indicators and shape parameter indicators.

[0175] Step S302: The needle-shaped tea detection terminal device performs local preprocessing on the various numerical detection commands of the needle-shaped tea to obtain the humidity detection result of the needle-shaped tea at the current moment.

[0176] Here, the humidity detection result includes detection parameters that correspond one-to-one with the index. These detection parameters are normalized values, which are obtained by preprocessing the numerical detection instructions of the corresponding index. The normalized values ​​are between [0,1].

[0177] Step S303: The needle-shaped tea detection terminal device queries the processing range value and the needle-shaped tea reference humidity parameter value that match the preset information of the needle-shaped tea locally;

[0178] Step S304: The needle-shaped tea detection device determines whether the current humidity state of the needle-shaped tea is dry based on the processing range value of the needle-shaped tea, the reference humidity parameter value of the needle-shaped tea, and the humidity detection result at the current moment. If it is dry, continue to step S305; otherwise, continue to the next moment of processing and return to step S301.

[0179] Step S305: The needle-shaped tea detection device adjusts the drying humidity and shape of the needle-shaped tea.

[0180] Step S306: Determine whether the needle-shaped tea detection device is currently online;

[0181] Step S307: If offline, the needle-shaped tea detection device locally stores the humidity detection result of the needle-shaped tea at the current moment.

[0182] Step S308: If online, the needle-shaped tea detection device uploads the humidity detection results of the needle-shaped tea image to be detected stored locally and its preset information to the cloud server through its front-end communication module.

[0183] In one implementation, the moisture content indicators of needle-shaped tea include, but are not limited to, respiration rate (breaths / minute), mass (mg), and fermentation rate variability. The shape parameters of needle-shaped tea include, but are not limited to, needle-shaped tea leaf curling rate (rad / s), needle-shaped tea leaf area (square centimeters), and needle-shaped tea leaf curl radius.

[0184] In one implementation, the local preprocessing in step S302 may include data preprocessing (e.g., data cleaning), unified sampling, data normalization, etc. The corresponding humidity detection result includes detection parameters that correspond one-to-one with the index. Each detection parameter includes a normalized value, which is obtained by local preprocessing through the numerical detection instruction of the corresponding index. The normalized value is between [0,1].

[0185] In one implementation, step S309 may include information verification and database matching steps. Here, information verification may include verifying whether the preset information of the needle-shaped tea is legal (e.g., whether there is a valid tea drinking certificate) or exists. The database matching step may include finding a humidity detection result library corresponding to the preset information of the needle-shaped tea. After finding the humidity detection result library of the needle-shaped tea image to be detected, the humidity detection result of the needle-shaped tea image to be detected can be stored in its humidity detection result library.

[0186] In one implementation, the humidity detection results of the needle-shaped tea images to be detected can be stored in the form of a time data sequence in the humidity detection result library. Each record in the time data sequence contains the time information of a humidity detection result and the normalized value of each tea drinking index corresponding to the humidity detection result.

[0187] In one example, the above time data sequence can be represented as shown in Table 1 below, where a, b, c, d, e, and f correspond to respiration frequency, mass, fermentation rate variability, needle-shaped tea leaf curling rate, needle-shaped tea leaf area, and needle-shaped tea leaf curling degree, respectively. In the humidity detection results at the i-th second (i = 1, 2, 3…), ai, bi, ci, di, ei, and fi represent the normalized values ​​of respiration frequency, mass, fermentation rate variability, needle-shaped tea leaf curling rate, needle-shaped tea leaf area, and needle-shaped tea leaf curling degree, respectively, at the i-th second. The fermentation rate variability is calculated based on oxygen concentration or fermentation rate.

[0188] Time (i-th second, i = 1, 2, 3...) Humidity detection result

[0189] 1 [a1,b1,c1,d1,e1,f1,…]

[0190] 2 [a2,b2,c2,d2,e2,f2,…]

[0191] 3 [a3,b3,c3,d3,e3,f3,…]

[0192] …… ……

[0193] In this embodiment, multi-dimensional humidity detection results are obtained through numerical detection commands of multiple indicators, and the humidity pattern of needle-shaped tea is determined in real time based on the multi-dimensional humidity detection results. It has low requirements for the collection environment, small pixel count, and low requirements for computing power and bandwidth for data processing and transmission, resulting in a better experience for needle-shaped tea and lower usage costs.

[0194] Figure 4 An exemplary flowchart of clustering in this application embodiment is provided. The exemplary clustering process in this application embodiment may include the following steps:

[0195] Step S401: Create two sets of needle-shaped tea images, namely K1, which is to be processed, and K2, which is to be dried needle-shaped tea images. K1 corresponds to the humidity morphology to be processed, and K2 corresponds to the dry humidity morphology.

[0196] Step S402: Randomly select two humidity detection results from the unclassified humidity detection results in the humidity detection result library of the needle-shaped tea images to be detected as the needle-shaped tea reference humidity parameter value X1 for the needle-shaped tea image set K1 to be processed and the needle-shaped tea reference humidity parameter value X2 for the dry needle-shaped tea image set K2.

[0197] Step S403: Calculate the difference between each humidity detection result and the reference humidity parameter value X1 and the reference humidity parameter value X2 of needle tea in the existing unclassified humidity detection results;

[0198] In one implementation, the difference can be calculated using the following equations (3) and (4):

[0199] D1=|ax-ax1|+|bx-bx1|+|cx-cx1|+|dx-dx1|+|ex-ex1|+|fx-fx1|+…(3)

[0200] D2=|ax-ax2|+|bx-bx2|+|cx-cx2|+|dx-dx2|+|ex-ex2|+|fx-fx2|+…(4)

[0201] Wherein, D1 represents the difference between the humidity detection result {ax, bx, cx, dx, ex, ...} and the first needle-shaped tea reference humidity parameter value X1{ax1, bx1, cx1, dx1, ex1, ...}, D2 represents the difference between the humidity detection result {ax, bx, cx, dx, ex, ...} and the second needle-shaped tea reference humidity parameter value X2{ax2, bx2, cx2, dx2, ex2, ...}, ax represents the normalized value of the corresponding tea drinking index a (e.g., respiratory rate) in the humidity detection result, ax1 represents the normalized value of the corresponding tea drinking index a (e.g., respiratory rate) in the needle-shaped tea reference humidity parameter value X1, ax2 represents the normalized value of the corresponding tea drinking index a (e.g., respiratory rate) in the needle-shaped tea reference humidity parameter value X2, and so on.

[0202] Step S404: Determine whether D1 is less than D2 to determine which needle tea image set the humidity detection result belongs to in the image set K1 of the needle tea to be processed and the image set K2 of the dried needle tea. If it is, continue to step S405; otherwise, continue to step S407.

[0203] Step S405: The humidity detection results are assigned to the needle-shaped tea image set K1 to be processed;

[0204] Step S406: Recalculate the mean number of the needle-shaped tea image set K1 to be processed, and use the mean number of the needle-shaped tea image set K1 to be processed as the new needle-shaped tea reference humidity parameter value X1', and continue to step S409.

[0205] Step S407: The humidity detection results are assigned to the dry needle-shaped tea image set K2;

[0206] Step S408: Recalculate the mean number of the dried needle-shaped tea image set K2 as the new reference humidity parameter value X2' for the needle-shaped tea;

[0207] Step S409: Determine whether the new needle-shaped tea reference humidity parameter value overlaps with the original needle-shaped tea reference humidity parameter value, that is, whether the new needle-shaped tea reference humidity parameter value X1' of the needle-shaped tea image set K1 to be processed overlaps with its original needle-shaped tea reference humidity parameter value X1, and whether the new needle-shaped tea reference humidity parameter value X2' of the dried needle-shaped tea image set K1 overlaps with its original needle-shaped tea reference humidity parameter value X2. If they overlap, the clustering ends, and the needle-shaped tea image set to be processed and its needle-shaped tea reference humidity parameter value, and the dried needle-shaped tea image set and its needle-shaped tea reference humidity parameter value are output. If any needle-shaped tea reference humidity parameter value does not overlap, continue to step S410.

[0208] In step S410, if any reference humidity parameter value of the needle-shaped tea does not overlap, the value is replaced, x1 = x1', x2 = x2', and the process returns to step S403 to cluster again until the mean center of the needle-shaped tea image set no longer moves.

[0209] In at least some embodiments, obtaining two sets of needle-shaped tea images through clustering can be represented by the following equations (5) and (6):

[0210] K1 = {xk11, xk12, xk13, ..., xk1j, ...}, where xk1j = [ak1j, bk1j, ck1j, dk1j, ek1j, fk1j, ...] (5)

[0211] K1 = {xk21, xk22, xk23, ..., xk2j, ...}, where xk2j = [ak2j, bk2j, ck2j, dk2j, ek2j, fk2j, ...] (6)

[0212] Where xk1j represents the j-th humidity detection result in the image set K1 of the needle-shaped tea to be processed, ak1j represents the normalized value of the corresponding tea drinking index a (e.g., respiratory rate) in the humidity detection result xk1j, xk2j represents the j-th humidity detection result in the image set K2 of the dried needle-shaped tea, ak2j represents the normalized value of the corresponding tea drinking index a (e.g., respiratory rate) in the humidity detection result xk2j, and so on, without further details.

[0213] Figure 4 The exemplary process is the initial clustering process, that is, the clustering process performed when neither the image set of needle-shaped tea to be processed nor the image set of dried needle-shaped tea exists. It can be understood that if both the image set of needle-shaped tea to be processed and the image set of dried needle-shaped tea exist, the clustering is based on the results of the previous clustering, that is, based on the image set of needle-shaped tea to be processed K1, the image set of dried needle-shaped tea K2, the reference humidity parameter value X1 of needle-shaped tea, and the reference humidity parameter value X2 of needle-shaped tea obtained from the previous clustering, from step S403 to step S411.

[0214] This embodiment uses a multidimensional data clustering method, a non-classification method, which avoids the problem of linear inseparability when the data dimension is high. Furthermore, the clustering results do not directly affect the classification of intermediate features such as brightness and color. Instead, the humidity morphology of the needle-shaped tea is directly characterized by the clustered needle-shaped tea image set. In this way, the humidity morphology of the needle-shaped tea can be directly determined by using the processing range value of the clustered needle-shaped tea image set and the reference humidity parameter value of the needle-shaped tea, without relying on intermediate features such as color and brightness. This reduces computational complexity, decreases false or missed identification, and improves the accuracy of needle-shaped tea humidity morphology detection.

[0215] Figure 5 An exemplary flowchart for configuring data filtering is shown. The exemplary process for configuring data filtering in this embodiment may include the following steps:

[0216] Step S501: Slice all humidity detection results in the needle-shaped tea image set into equal-range pixel slices to obtain multiple slices with the same number of pixels.

[0217] For example, the needle-shaped tea image set is sliced ​​into 100 pieces, and the range of each slice is 0.01 pixels.

[0218] Step S502: Sort the slices obtained by dividing the pixel grid according to the number of pixels and find the slice with the largest number of pixels.

[0219] Assume the range of the slice with the largest number of pixels is [u, u+0.01], and the data range of the needle-shaped tea image set is (0, 1).

[0220] Step S503: Determine whether the number of pixels in the first range and the number of pixels in the second range of the needle-shaped tea image are equal. If they are equal, continue to step S504; otherwise, proceed to step S505.

[0221] Step S504: Determine the data configuration of the needle-shaped tea image set as a balanced configuration;

[0222] Step S505: Determine that the data configuration of the needle-shaped tea image set is an unbalanced configuration.

[0223] If the detection parameters in the humidity detection results have been normalized during local preprocessing, the value of each detection parameter in the humidity detection results will be configured between [0,1], and no further normalization is required before step S501 in this process. If the value of the detection parameter is not a normalized value configured between [0,1], then a step of normalizing each humidity detection result in the needle-shaped tea image set needs to be added before step S501.

[0224] In this embodiment, data configuration is filtered by dividing the data into pixel grids. Different processing range values ​​are calculated for different types of data configurations, thereby obtaining various processing range values ​​that can accurately characterize the boundaries of various humidity forms of needle-shaped tea, providing an accurate basis for the accurate detection of the humidity form of needle-shaped tea.

[0225] This embodiment provides a detailed description of an exemplary implementation method for adjusting the moisture content of needle-shaped tea.

[0226] In some implementations, the triggering process for adjusting the humidity state of needle-shaped tea may include: determining whether the duration of the needle-shaped tea in a dry humidity state has reached a pre-set second predetermined value; if so, then triggering the adjustment of the humidity state of the needle-shaped tea; otherwise, no intervention is needed. This implementation adjusts the humidity state of the needle-shaped tea only after it has been in a dry humidity state for a certain period of time, which can avoid false triggering of the adjustment mechanism and improve the needle-shaped tea experience.

[0227] Figure 6 An exemplary implementation process for adjusting the moisture content of needle-shaped tea is shown. The exemplary implementation process for adjusting the moisture content of needle-shaped tea may include the following steps:

[0228] Step S601: Determine the degree of dryness of the needle-shaped tea in terms of its moisture content.

[0229] Step S602: Adjust the humidity of the needle-shaped tea according to the preset three-level adjustment mechanism. After the predetermined third time, redetermine the dryness of the needle-shaped tea humidity. If the dryness decreases, continue to step S603. If the dryness is still at a high level, jump to step S605.

[0230] Step S603: Adjust the humidity state of the needle-shaped tea according to the preset two-level adjustment mechanism. After the predetermined second time, redetermine the dryness of the humidity state of the needle-shaped tea. If the dryness decreases, continue to step S604. If the dryness is still at the medium level or the dryness increases to the high level, jump to step S602.

[0231] Step S604: Adjust the humidity state of the needle-shaped tea according to the preset two-level adjustment mechanism. After the predetermined second time, redetermine the dryness of the humidity state of the needle-shaped tea. If the dryness decreases, continue to step S604. If the dryness is still at the medium level or the dryness increases to the high level, jump to step S602.

[0232] Step S605: Adjust the humidity of the needle-shaped tea according to the preset four-level adjustment mechanism. After the predetermined third time period, redetermine the dryness of the needle-shaped tea's humidity. If the dryness decreases, return to step S603. If the dryness is still at a high level, this step can be repeated.

[0233] Generally, the higher the degree of dryness, the longer the time required to regulate the state of the needle-shaped tea. Therefore, in at least some embodiments, a stepped time threshold can be used, where the first time is less than the second time, and the second time is less than the third time. The specific values ​​of the first, second, and third times can be empirical values ​​or dynamically changing values, and can be adjusted according to the characteristics of the needle-shaped tea.

[0234] In some implementations, determining the degree of dryness can involve pre-dividing each processing range value within a category of processing range values ​​corresponding to a specific dryness level into multiple sub-ranges. These sub-ranges represent different levels of dryness. The current dryness level of the needle-shaped tea is determined by identifying which sub-range the average value of the measured parameter from all humidity measurements within the current adjustment period falls into. For example, each processing range value within a category of processing range values ​​corresponding to a specific dryness level can be equally divided into three sub-ranges, representing high, medium, and low dryness levels, respectively. If the average value of the same measured parameter from all humidity measurements within the first time period falls into the corresponding sub-range, then the current dryness level of the needle-shaped tea is determined to be the level corresponding to that sub-range.

[0235] In some implementation methods, Figure 6 The diagram illustrates the structural features of each level of regulatory mechanism. A first-level regulatory mechanism can be a single-mode intervention, using one detection method (e.g., mode 1) to intervene in the needle-shaped tea state; a second-level regulatory mechanism can be a dual-mode intervention, using two detection methods (e.g., mode 1 and mode 2) to intervene in the needle-shaped tea state; a third-level regulatory mechanism can be a three-mode intervention, using three detection methods (e.g., mode 1, mode 2, and mode 3) to intervene in the needle-shaped tea state; and a fourth-level regulatory mechanism can be a four-mode intervention, using four detection methods (e.g., mode 1, mode 2, mode 3, and mode 4) to intervene in the needle-shaped tea state. It should be noted that the numbers "1-4" in mode 1, mode 2, mode 3, and mode 4 are only used to distinguish the mode type, not to limit the specific mode.

[0236] In some implementation methods, the adjustment mechanism is randomly determined. That is, one or more of the following four methods are randomly selected: re-rolling, initial rolling, second fermentation, and hot air dehydration. By adjusting the moisture content of needle-shaped tea through random selection, the tolerance of the tea caused by prolonged use of the same detection instructions can be effectively prevented, and the tolerance resistance of the adjustment mechanism can be significantly improved.

[0237] The adjustment method in this embodiment has the function of adjusting resistance.

[0238] In this embodiment, when the newly generated humidity detection result falls into a certain value, the strategy for entering the humidity detection result database is changed from directly sending the humidity detection result into the humidity detection result database to "only selecting points that do not fall into the processing range value dimension greater than half to enter the database". As a result, while continuously improving the data image of the needle-shaped tea state, redundant data entering the humidity detection result database can be reduced, thereby reducing the amount of data processing and lowering the computational complexity.

[0239] In the system provided in this embodiment, the processing range value and reference humidity parameter value of needle-shaped tea can be dynamically adjusted as the humidity detection results of needle-shaped tea are updated. This not only improves the detection accuracy of the humidity morphology of needle-shaped tea, but also allows for dynamic detection of the humidity morphology of needle-shaped tea according to the specific type of needle-shaped tea and varies from time to time, which is more in line with the actual situation of needle-shaped tea consumption.

[0240] Those skilled in the art will recognize that the apparatus and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Skilled tea enthusiasts may use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0241] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and apparatuses described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0242] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of apparatus is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple apparatuses or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or processing connection shown or discussed may be through some interfaces, or indirect coupling or processing connection between apparatuses, and may be electrical, mechanical, or other forms.

[0243] The devices described as separate components may or may not be physically separate. The components shown as devices may or may not be physical devices; that is, they may be located in one place or configured on multiple network devices. Some or all of the devices can be selected to achieve the purpose of this embodiment according to actual needs.

[0244] In addition, the functional devices in the various embodiments of this application can be integrated into a processing device, or each device can exist physically separately, or two or more devices can be integrated into a device.

[0245] If a function is implemented as a software device and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which can be a pin-type computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0246] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, performs a method for detecting the humidity of needle-shaped tea based on image recognition, the method including at least one of the schemes described in the above embodiments.

[0247] The computer storage medium in this application embodiment can be any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable detection instruction medium or a computer-readable storage medium. For example, a computer-readable storage medium can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory, a read-only memory, an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0248] Note that the above are merely preferred embodiments and the technical principles employed in this application. Those skilled in the art will understand that this invention is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions can be made without departing from the scope of protection of this invention. Therefore, although this application has been described in detail through the above embodiments, this invention is not limited to the above embodiments. Many other equivalent embodiments may be included without departing from the concept of this invention, all of which fall within the scope of protection of this invention.

Claims

1. An image recognition-based needle-shaped tea moisture detection method, characterized by, The method comprises the following steps: obtaining a humidity detection result of a to-be-detected image of needle-shaped tea, the humidity detection result comprising at least two detection parameters; when the number of to-be-detected images of the detection parameter falling within the fixation treatment range value is greater than a first predetermined value among the at least two detection parameters, determining the humidity form corresponding to the fixation treatment range value as the humidity form of the needle-shaped tea; the at least two detection parameters are used to indicate the water content and / or shape parameter; the humidity form comprises a to-be-processed humidity form or a dry humidity form, and the fixation treatment range value corresponds to the to-be-processed humidity form or the dry humidity form; the fixation treatment range value is obtained by the following method: performing clustering based on a humidity detection result library of to-be-detected images of needle-shaped tea to form a set of needle-shaped tea images corresponding to the humidity form, the humidity detection result library of to-be-detected images of needle-shaped tea comprising pre-obtained humidity detection results of to-be-detected images of needle-shaped tea; performing data configuration screening on the set of needle-shaped tea images to determine whether the data configuration of the set of needle-shaped tea images belongs to a balanced configuration or an unbalanced configuration; determining the fixation treatment range value corresponding to the humidity form according to the data configuration of the set of needle-shaped tea images; the data configuration screening on the set of needle-shaped tea images specifically comprises: dividing the humidity detection results in the set of needle-shaped tea images into a plurality of pixel grids to obtain a plurality of pixel grids, the number of pixels in each pixel grid being the same; taking the pixel grid with the largest number of pixels in the plurality of pixel grids as a mean value range of the set of needle-shaped tea images; when the number of pixels of the set of needle-shaped tea images in a first range is equal to the number of pixels of the set of needle-shaped tea images in a second range, the data configuration of the set of needle-shaped tea images is a balanced configuration; when the number of pixels of the set of needle-shaped tea images in the first range is not equal to the number of pixels of the set of needle-shaped tea images in the second range, the data configuration of the set of needle-shaped tea images is an unbalanced configuration; wherein the first range and the second range are obtained by dividing the data range of the set of needle-shaped tea images according to the mean value range.

2. The image recognition-based needle tea moisture detection method according to claim 1, characterized in that, The method further comprises: when the number of to-be-detected images of the detection parameter falling within the fixation treatment range value is less than or equal to the first predetermined value among the at least two detection parameters, determining the humidity form of the needle-shaped tea according to the difference between the humidity detection result and a first reference humidity parameter value of the needle-shaped tea; the determination of the humidity form of the needle-shaped tea according to the difference between the humidity detection result and the first reference humidity parameter value of the needle-shaped tea comprises: among two first reference humidity parameter values of the needle-shaped tea, determining the humidity form corresponding to the first reference humidity parameter value closer to the humidity detection result as the humidity form of the needle-shaped tea; or among three or more first reference humidity parameter values of the needle-shaped tea, determining the humidity form corresponding to the first reference humidity parameter value closest to the humidity detection result as the humidity form of the needle-shaped tea; the determination of the humidity form of the needle-shaped tea according to the difference between the humidity detection result and the first reference humidity parameter value of the needle-shaped tea comprises: when an absolute value of a difference between the first difference value and the second difference value is greater than a mean value of the first difference value and the second difference value, the humidity form is determined as the humidity form to be processed; when the absolute value of the difference between the first difference value and the second difference value is less than or equal to the mean value of the first difference value and the second difference value, the humidity form is determined as the dry humidity form; the first difference value is a difference between the humidity detection result and a first needle-shaped tea reference humidity parameter value corresponding to the humidity form to be processed, and the second difference value is a difference between the humidity detection result and a first needle-shaped tea reference humidity parameter value corresponding to the dry humidity form.

3. The image recognition-based needle tea moisture detection method according to claim 2, characterized in that, Further comprising: when a humidity detection result falling rate of the needle-shaped tea image to be detected is greater than or equal to a third predetermined value, adding a humidity detection result of the needle-shaped tea image to be detected falling into the detection parameter of the fixation processing range value and less than or equal to the first predetermined value to a humidity detection result library of the needle-shaped tea image to be detected; when a duration of the dry humidity form is greater than or equal to a second predetermined value, adjusting the humidity form by using one or more adjustment mechanisms including the following ways: a primary rolling way, a secondary rolling way, a double-fermentation way, and a hot air dehydration way; the adjustment mechanism is determined according to a drying priority of the needle-shaped tea, and the drying priority is determined according to the humidity detection result and a fixation processing range value corresponding to the humidity form; redetermining a first drying priority of the needle-shaped tea after a first time interval corresponding to the first drying priority; the way in the adjustment mechanism is randomly selected.

4. An image recognition-based needle-shaped tea moisture detection system, characterized by, Comprising: an acquisition device configured to acquire a humidity detection result of a needle-shaped tea image to be detected, the humidity detection result including at least two detection parameters; a determination device configured to determine a humidity form corresponding to a fixation processing range value as a humidity form of the needle-shaped tea when a number of images to be detected falling into the detection parameter of the fixation processing range value is greater than a first predetermined value in the at least two detection parameters; the at least two detection parameters are used to indicate a water content and / or a shape parameter; the humidity form includes a humidity form to be processed and a dry humidity form, and the fixation processing range value corresponds to the humidity form to be processed or the dry humidity form; a calculation device configured to: cluster a humidity detection result library of a needle-shaped tea image to be detected to form a needle-shaped tea image set corresponding to the humidity form, the humidity detection result library of the needle-shaped tea image to be detected including a humidity detection result of a needle-shaped tea image to be detected obtained in advance; perform data configuration screening on the needle-shaped tea image set to determine whether a data configuration of the needle-shaped tea image set belongs to a balanced configuration or an unbalanced configuration; determine a fixation processing range value corresponding to the humidity form according to the data configuration of the needle-shaped tea image set; the calculation device is specifically configured to: perform pixel grid division on the humidity detection result in the needle-shaped tea image set to obtain a plurality of pixel grids, and a number of pixels in each pixel grid in the plurality of pixel grids is the same; take a pixel grid with the largest number of pixels in the plurality of pixel grids as a mean value range of the needle-shaped tea image set; the needle-shaped tea image set is configured in a balanced configuration when the number of pixels in the first range is equal to the number of pixels in the second range; the needle-shaped tea image set is configured in a non-balanced configuration when the number of pixels in the first range is not equal to the number of pixels in the second range; wherein the first range and the second range are obtained by dividing the data range of the needle-shaped tea image set according to the mean range.

5. The image recognition based needle tea moisture detection system according to claim 4, wherein, The determination device is further configured to determine the humidity form of the needle-shaped tea according to the difference between the humidity detection result and a first needle-shaped tea reference humidity parameter value when the number of images to be detected in the detection parameter falling within the range of the fixation treatment value is less than or equal to the first predetermined value among the at least two detection parameters. The determination device is specifically configured to: determine the humidity form corresponding to the first needle-shaped tea reference humidity parameter value that is closer to the humidity detection result as the humidity form of the needle-shaped tea among the two first needle-shaped tea reference humidity parameter values; or, determine the humidity form corresponding to the first needle-shaped tea reference humidity parameter value that is closest to the humidity detection result as the humidity form of the needle-shaped tea among the three or more first needle-shaped tea reference humidity parameter values. The determination device is specifically configured to: determine the humidity form as a humidity form to be processed when the absolute value of the difference between the first difference value and the second difference value is greater than the mean value of the first difference value and the second difference value; determine the humidity form as a dry humidity form when the absolute value of the difference between the first difference value and the second difference value is less than or equal to the mean value of the first difference value and the second difference value; the first difference value is the difference between the humidity detection result and the first needle-shaped tea reference humidity parameter value corresponding to the humidity form to be processed, and the second difference value is the difference between the humidity detection result and the first needle-shaped tea reference humidity parameter value corresponding to the dry humidity form.

6. The image recognition based needle tea moisture detection system according to claim 5, wherein, Further comprising: a database updating device configured to add the humidity detection result, in which the number of images to be detected in the detection parameter falling within the range of the fixation treatment value is less than or equal to the first predetermined value, to a humidity detection result database of images to be detected of the needle-shaped tea when the humidity detection result of the images to be detected of the needle-shaped tea falls within a rate greater than or equal to a third predetermined value; an adjusting device configured to adopt one or more adjusting mechanisms including the following ways to adjust the humidity form when the duration of the dry humidity form is greater than or equal to a second predetermined value: a primary rolling way, a secondary rolling way, a double-fixation way, and a hot-air dehydration way; the adjusting mechanism is determined according to the drying priority of the needle-shaped tea, and the drying priority is determined according to the humidity detection result and the fixation treatment range value corresponding to the humidity form; the adjusting device is further configured to re-determine the first drying priority of the needle-shaped tea after a first time interval corresponding to the first drying priority; the way in the adjusting mechanism is randomly selected by the adjusting device.

Images