An electronic cigarette preparation method and system based on image analysis
By using image analysis technology and constructing a knowledge graph with variational autoencoders and long short-term neural networks, the problem of inaccurate oil filling speed in electronic cigarette atomizers was solved, achieving uniform oil filling and efficient production.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- 苑卉
- Filing Date
- 2026-04-03
- Publication Date
- 2026-07-21
AI Technical Summary
In existing technologies, the e-liquid filling process of electronic cigarette atomizers is difficult to control precisely, resulting in uneven e-liquid distribution, which affects the atomization effect and taste. Furthermore, high-speed filling can easily introduce air bubbles, while low-speed filling reduces production efficiency.
An image-based analysis method is used to acquire X-ray images of the atomizer at different e-liquid injection speeds. A variational autoencoder is used to generate e-liquid injection videos. A knowledge graph is constructed by combining a long short-term neural network and a graph autoencoder to determine the target e-liquid injection speed and achieve precise e-liquid injection.
It achieves precise control of the atomizer's e-liquid filling speed, improves atomization effect and flavor consistency, optimizes the production process, and enhances production efficiency and product quality.
Smart Images
Figure CN121970936B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of electronic cigarette manufacturing technology, specifically to an electronic cigarette manufacturing method and system based on image analysis. Background Technology
[0002] In recent years, with increasing health awareness and restrictions on traditional tobacco products, e-cigarettes have rapidly emerged as an alternative. However, the manufacturing process of e-cigarettes still faces many challenges, especially in the e-liquid filling stage inside the atomizer. Traditional filling methods rely mainly on experience or simple mechanical control, making it difficult to precisely adjust the filling speed. Rapid filling can lead to uneven distribution of e-liquid within the atomizer, resulting in some areas being overly wet while others are dry. This unevenness affects the efficiency of the heating element, thus impacting the atomization effect and flavor. High-speed filling can easily introduce air, forming bubbles. These bubbles affect the fluidity of the e-liquid and may cause intermittent problems during atomization, such as a "dry burning" sensation or a deterioration in flavor. Conversely, excessively slow filling speeds significantly reduce the overall efficiency of the production line, prolonging the production cycle of each product and increasing manufacturing costs.
[0003] Therefore, accurately determining the e-liquid filling speed of the atomizer is a problem that urgently needs to be solved. Summary of the Invention
[0004] The main technical problem this invention addresses is how to accurately determine the oil injection speed of an atomizer.
[0005] According to a first aspect, the present invention provides an image analysis-based method for manufacturing an electronic cigarette, comprising: acquiring multiple consecutive X-ray images of an atomizer taken at different e-liquid injection speeds; generating an internal e-liquid filling video of the atomizer at each e-liquid injection speed using a variational autoencoder based on the multiple consecutive X-ray images of the atomizer taken at each e-liquid injection speed; determining the internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds using an e-liquid filling information determination model based on the internal e-liquid filling video of the atomizer at each e-liquid injection speed; determining a target e-liquid filling speed based on the internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds; and filling the electronic cigarette with e-liquid based on the target e-liquid filling speed.
[0006] In one possible implementation, the oil injection information determination model is a long short-term neural network model.
[0007] In one possible implementation, determining the target e-liquid injection speed based on the atomizer's internal e-liquid injection information at each e-liquid injection speed and the similarity of the atomizer's internal e-liquid injection information at different e-liquid injection speeds includes: A knowledge graph is constructed, which includes multiple nodes and multiple edges between the nodes. Each node represents an e-liquid injection speed, and the edges between nodes represent the similarity of the internal e-liquid injection information of the atomizer at the e-liquid injection speed. The node features of each node include the internal e-liquid injection information of the atomizer at that e-liquid injection speed. The target oil injection speed is determined by processing the knowledge graph using a graph autoencoder.
[0008] In one possible implementation, the internal e-liquid filling information of the atomizer at each e-liquid injection speed includes e-liquid uniformity, bubble formation information, e-liquid filling degree, and e-liquid sealing degree.
[0009] According to a second aspect, the present invention provides an image analysis-based electronic cigarette manufacturing system, comprising: The acquisition module is used to acquire multiple consecutive X-ray images of the atomizer at different e-liquid injection speeds; The generation module is used to generate an internal e-liquid filling video of the atomizer at each e-liquid filling speed using a variational autoencoder based on multiple consecutive X-ray images of the atomizer taken at each e-liquid filling speed. The information determination module is used to determine the internal e-liquid filling information of the atomizer at each e-liquid filling speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid filling speeds based on the internal e-liquid filling video of the atomizer at each e-liquid filling speed using the e-liquid filling information determination model. The e-liquid injection speed determination module is used to determine the target e-liquid injection speed based on the internal e-liquid injection information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid injection information of the atomizer at different e-liquid injection speeds. The e-liquid filling module is used to fill the e-cigarette with e-liquid based on the target e-liquid filling speed.
[0010] In one possible implementation, the oil injection information determination model is a long short-term neural network model.
[0011] In one possible implementation, determining the target e-liquid injection speed based on the atomizer's internal e-liquid injection information at each e-liquid injection speed and the similarity of the atomizer's internal e-liquid injection information at different e-liquid injection speeds includes: A knowledge graph is constructed, which includes multiple nodes and multiple edges between the nodes. Each node represents an e-liquid injection speed, and the edges between nodes represent the similarity of the internal e-liquid injection information of the atomizer at the e-liquid injection speed. The node features of each node include the internal e-liquid injection information of the atomizer at that e-liquid injection speed. The target oil injection speed is determined by processing the knowledge graph using a graph autoencoder.
[0012] In one possible implementation, the internal e-liquid filling information of the atomizer at each e-liquid injection speed includes e-liquid uniformity, bubble formation information, e-liquid filling degree, and e-liquid sealing degree.
[0013] According to a third aspect, embodiments of the present invention provide an electronic device, including: a processor; a memory; and a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the method described above, the method comprising: acquiring a series of consecutive X-ray images of an atomizer taken at different e-liquid injection speeds; generating an internal e-liquid filling video of the atomizer at each e-liquid injection speed using a variational autoencoder based on the series of consecutive X-ray images of the atomizer taken at each e-liquid injection speed; determining, based on the internal e-liquid filling video of the atomizer at each e-liquid injection speed, internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds using an e-liquid filling information determination model; determining a target e-liquid filling speed based on the internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds; and filling an electronic cigarette with e-liquid based on the target e-liquid filling speed.
[0014] According to the fourth aspect, this embodiment provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned image analysis-based electronic cigarette manufacturing method. The method includes: acquiring multiple consecutive X-ray images of an atomizer taken at different e-liquid injection speeds; generating an internal e-liquid filling video of the atomizer at each e-liquid injection speed using a variational autoencoder based on the multiple consecutive X-ray images of the atomizer taken at each e-liquid injection speed; determining the internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds using an e-liquid filling information determination model based on the internal e-liquid filling video of the atomizer at each e-liquid injection speed; determining a target e-liquid filling speed based on the internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds; and filling the electronic cigarette with e-liquid based on the target e-liquid filling speed.
[0015] This invention provides an image analysis-based method and system for manufacturing electronic cigarettes. The method includes acquiring multiple consecutive X-ray images of an atomizer at different e-liquid injection speeds; generating internal e-liquid filling videos of the atomizer at each injection speed using a variational autoencoder based on the multiple consecutive X-ray images; determining the internal e-liquid filling information of the atomizer at each injection speed and the similarity of the internal e-liquid filling information at different injection speeds using an e-liquid filling information determination model based on the internal e-liquid filling videos; determining a target e-liquid filling speed based on the internal e-liquid filling information of the atomizer at each injection speed and the similarity of the internal e-liquid filling information at different injection speeds; and filling the electronic cigarette with e-liquid based on the target e-liquid filling speed. This method can accurately determine the e-liquid filling speed of the atomizer. Attached Figure Description
[0016] Figure 1 A schematic diagram illustrating an application scenario of an image analysis-based electronic cigarette manufacturing method provided in an embodiment of the present invention; Figure 2 A schematic flowchart of an image analysis-based electronic cigarette manufacturing method provided in an embodiment of the present invention; Figure 3 A schematic diagram of a process for determining a target oil injection rate is provided in an embodiment of the present invention; Figure 4 A schematic diagram of an image analysis-based electronic cigarette manufacturing system provided in an embodiment of the present invention; Figure 5 A schematic diagram of an electronic device provided in an embodiment of the present invention; Figure 6 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation
[0017] The present invention will now be described in further detail with reference to specific embodiments and accompanying drawings. Similar elements in different embodiments are referred to by associated similar element reference numerals. In the following embodiments, many details are described to facilitate a better understanding of the invention. However, those skilled in the art will readily recognize that some features may be omitted in different situations, or may be replaced by other elements, materials, or methods. In some cases, certain operations related to the present invention are not shown or described in the specification. This is to avoid obscuring the core parts of the invention with excessive description. For those skilled in the art, detailed description of these related operations is not necessary; they can fully understand the related operations based on the description in the specification and general technical knowledge in the art.
[0018] Figure 1 This is a schematic diagram illustrating an application scenario of an image analysis-based electronic cigarette manufacturing method provided in an embodiment of the present invention. Figure 1 The application scenarios of the image analysis-based electronic cigarette manufacturing method can include servers 11, networks 12, terminals 13, and storage devices 14.
[0019] In some embodiments, server 11 may be a single server or a group of servers. Server 11 can access information and / or data stored in terminal 13 or storage device 14 via network 12. In some embodiments, server 11 may be used to perform... Figure 2 The image analysis-based electronic cigarette manufacturing method shown is illustrated.
[0020] Network 12 can facilitate the exchange of information and / or data. In some embodiments, network 12 can be any form of wired or wireless network, or any combination thereof.
[0021] Terminal 13 may refer to one or more terminal devices used by a user. In some embodiments, terminal 13 may include one or more combinations of mobile devices, tablet computers, laptop computers, etc.
[0022] Storage device 14 can store data and / or instructions, for example, storage device 14 can store data instructions for an electronic cigarette manufacturing method based on image analysis.
[0023] In this embodiment of the invention, the following are provided: Figure 2 The image analysis-based electronic cigarette manufacturing method shown includes steps S1 to S5: Step S1: Obtain multiple consecutive X-ray images of the atomizer taken at different e-liquid injection speeds.
[0024] E-liquid fill speed refers to the rate at which e-liquid enters the atomizer during the filling process. Different speeds can affect the filling effect. E-liquid fill speeds can be categorized as slow, medium, and fast.
[0025] A series of consecutive X-ray images taken of an atomizer at different e-liquid infusion rates is a series of images obtained by continuously photographing an electronic cigarette atomizer. These images can show the internal structure of the atomizer and how it changes over time.
[0026] As an example, three different e-liquid injection speeds were used, and 10 consecutive X-ray images were taken at each speed to observe the changes inside the atomizer.
[0027] Step S2: Based on multiple consecutive X-ray images of the atomizer taken at each e-liquid injection speed, a variational autoencoder is used to generate an internal e-liquid injection video of the atomizer at each e-liquid injection speed.
[0028] The internal e-liquid filling video of the atomizer at each e-liquid injection speed is a dynamic video generated based on a variational autoencoder, showing how e-liquid enters and fills the atomizer at a specific e-liquid injection speed.
[0029] A variational autoencoder (VAE) is a generative model used to learn latent representations of data and generate new data samples. A VAE consists of an encoder and a decoder. The encoder maps the input data to a latent space, and the decoder maps points in the latent space back to the data space.
[0030] The input to the variational autoencoder is a series of X-ray images of the atomizer taken at each e-liquid injection speed, and the output of the variational autoencoder is an internal e-liquid injection video of the atomizer at each e-liquid injection speed.
[0031] The encoder of a variational autoencoder (VAE) can extract key features, such as e-liquid flow path and filling degree, from multiple consecutive X-ray images. These features are mapped into a latent space, forming a low-dimensional, compact representation. Each point in the latent space not only represents a specific set of input images but also implies the relationships and patterns of change between these images. For example, different e-liquid injection speeds may create different regions or trajectories in the latent space. The decoder of a VAE can generate a corresponding output from any point in the latent space, meaning that dynamic videos can be generated based on the representation in the latent space, showcasing different stages of the e-liquid injection process.
[0032] Step S3: Based on the internal e-liquid injection video of the atomizer at each e-liquid injection speed, use the e-liquid injection information determination model to determine the internal e-liquid injection information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid injection information of the atomizer at different e-liquid injection speeds.
[0033] The e-liquid filling information determination model is a long short-term neural network model. The input of the e-liquid filling information determination model is the internal e-liquid filling video of the atomizer at each e-liquid filling speed, and the output of the e-liquid filling information determination model is the internal e-liquid filling information of the atomizer at each e-liquid filling speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid filling speeds.
[0034] Long Short-Term Memory (LSTM) neural network models can process the internal e-liquid filling video of the atomizer at each e-liquid filling speed over a continuous time period. This allows for better capture of the relationships within the time series of the internal e-liquid filling video at each filling speed, outputting features that comprehensively consider the correlations between the internal e-liquid filling video at each time point and each e-liquid filling speed. This results in more accurate and comprehensive output features.
[0035] The internal e-liquid filling information of the atomizer at each e-liquid injection speed includes e-liquid uniformity, bubble formation information, e-liquid filling degree, and e-liquid sealing degree.
[0036] E-liquid uniformity refers to the degree of evenness in the distribution of e-liquid within the atomizer.
[0037] Bubble formation information indicates whether bubbles were formed during the oil injection process and their distribution. For example, bubble formation information may indicate no obvious bubbles.
[0038] The filling degree refers to the proportion or extent to which e-liquid fills the atomizer.
[0039] The degree of sealing after oil filling refers to the sealing condition of the atomizer after oil filling.
[0040] In some embodiments, the e-liquid filling information determination model includes an e-liquid filling information determination layer and a similarity determination layer. The input to the e-liquid filling information determination layer is the internal e-liquid filling video of the atomizer at each e-liquid filling speed, and the output of the e-liquid filling information determination layer is the internal e-liquid filling information of the atomizer at each e-liquid filling speed. The input to the similarity determination layer is the internal e-liquid filling information of the atomizer at each e-liquid filling speed, and the output of the similarity determination layer is the similarity of the internal e-liquid filling information of the atomizer at different e-liquid filling speeds.
[0041] Step S4: Determine the target e-liquid injection speed based on the internal e-liquid injection information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid injection information of the atomizer at different e-liquid injection speeds.
[0042] In some embodiments, Figure 3 This is a schematic diagram of a process for determining a target oil injection rate according to an embodiment of the present invention. The determination of the target oil injection rate includes steps S21 to S22: Step S21: Construct a knowledge graph. The knowledge graph includes multiple nodes and multiple edges between the nodes. Each node represents an e-liquid injection speed. The edges between nodes represent the similarity of the internal e-liquid injection information of the atomizer at the e-liquid injection speed. The node features of each node include the internal e-liquid injection information of the atomizer at that e-liquid injection speed. A knowledge graph is a graphical data structure used to represent entities and the relationships between them. Each node represents an e-liquid injection speed, and the edges between nodes represent the similarity of the atomizer's internal e-liquid injection information at that injection speed. The node features of each node include the atomizer's internal e-liquid injection information at that injection speed.
[0043] Step S22: Process the knowledge graph based on the graph autoencoder to determine the target oil injection speed.
[0044] A graph autoencoder (GAE) is a deep learning model specifically designed for processing graph data. It can compress node features while preserving the graph structure and extract useful information from it. A graph autoencoder typically consists of two parts: an encoder and a decoder. The encoder maps the graph data to a low-dimensional space, and the decoder reconstructs the original graph data from the low-dimensional space.
[0045] Knowledge graphs can be used to represent complex oil injection information and its interrelationships in a structured way. This structured representation helps to more clearly understand the similarities and differences between different oil injection rates and provides a foundation for subsequent analysis.
[0046] Knowledge graphs not only display the characteristics of each node (i.e., each e-liquid pouring speed), but also reveal the similarities and differences between different nodes through edges. For example, slow and medium pouring speeds may be similar in some aspects but different in others. Through the graph structure, these similarities and differences can be visually observed, leading to a better understanding of the characteristics of each pouring speed.
[0047] After processing by the graph autoencoder, the optimal e-liquid injection speed can be selected, and the optimal e-liquid injection speed performs best in multiple dimensions.
[0048] By processing knowledge graphs through graph autoencoders, the target oiling speed that performs best across multiple dimensions can be identified, thereby optimizing the production process and improving product quality and consistency.
[0049] Step S5: Fill the electronic cigarette with e-liquid based on the target filling speed.
[0050] Once the target refill speed is determined, the e-cigarette is refilled based on that target refill speed.
[0051] Based on the same inventive concept Figure 4 This is a schematic diagram of an image analysis-based electronic cigarette manufacturing system provided in an embodiment of the present invention. The image analysis-based electronic cigarette manufacturing system includes: Module 41 is used to acquire screen recordings of the user's preferred software and descriptions of data transmission. Analysis module 42 is used to determine the user's preferred interface and user interface preference information based on the screen recording of the running video of the user's preferred software using an analysis model. The module quantity determination module 43 is used to determine the minimum number of modules required by the user and the maximum number of modules required by the user based on the data transmission description text. The generation module 44 is used to generate multiple simulated software interaction interfaces using a variational autoencoder based on the number of modules required by the user at the lowest level, the number of modules required by the user at the highest level, the user's preferred software interaction interface, and the user interface preference information. The interface determination module 45 is used to determine the target software interaction interface based on the plurality of simulated software interaction interfaces.
[0052] Based on the same inventive concept, embodiments of the present invention provide an electronic device, such as... Figure 5 As shown, it includes: The system includes: a processor 51; a memory 52; and a computer program; wherein the computer program is stored in the memory 52 and configured to be executed by the processor 51 to implement the image analysis-based electronic cigarette manufacturing method provided above, the method comprising: acquiring multiple consecutive X-ray images of an atomizer taken at different e-liquid injection speeds; generating an internal e-liquid filling video of the atomizer at each e-liquid injection speed using a variational autoencoder based on the multiple consecutive X-ray images of the atomizer taken at each e-liquid injection speed; determining the internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds using an e-liquid filling information determination model based on the internal e-liquid filling video of the atomizer at each e-liquid injection speed; determining a target e-liquid filling speed based on the internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds; and filling the electronic cigarette with e-liquid based on the target e-liquid filling speed.
[0053] Based on the same inventive concept, this embodiment provides a computer-readable storage medium storing a computer program. When executed by processor 51, the program implements the aforementioned image analysis-based electronic cigarette manufacturing method. The method includes: acquiring multiple consecutive X-ray images of an atomizer taken at different e-liquid injection speeds; generating an internal e-liquid filling video of the atomizer at each e-liquid injection speed using a variational autoencoder based on the multiple consecutive X-ray images of the atomizer taken at each e-liquid injection speed; determining the internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds using an e-liquid filling information determination model based on the internal e-liquid filling video of the atomizer at each e-liquid injection speed; determining a target e-liquid filling speed based on the internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds; and filling the electronic cigarette with e-liquid based on the target e-liquid filling speed.
[0054] The image analysis-based electronic cigarette manufacturing method provided in this application can be applied to terminal devices (such as mobile phones), tablets, laptops, ultra-mobile personal computers (UMPCs), handheld computers, netbooks, personal digital assistants (PDAs), wearable devices (such as smartwatches, smart glasses, or smart helmets), augmented reality (AR) / virtual reality (VR) devices, smart home devices, in-vehicle computers, and other electronic devices. This application does not impose any limitations on this.
[0055] Taking mobile phone 100 as an example of the aforementioned electronic devices, Figure 6 A structural schematic diagram of mobile phone 100 is shown.
[0056] like Figure 6 As shown, the mobile phone 100 may include a processing module 110, an external memory interface 120, an internal memory 121, a universal serial bus (USB) interface 130, a charging management module 140, a power management module 141, a battery 142, an antenna 1, an antenna 2, a mobile communication module 150, a wireless communication module 160, an audio module 170, a speaker 170A, a receiver 170B, a microphone 170C, a headphone jack 170D, a sensor module 180, buttons 190, a motor 191, an indicator 192, a camera 193, a display screen 194, and a subscriber identification module (SIM) card interface 195, etc.
[0057] The processing module 110 can be used to: acquire multiple consecutive X-ray images of the atomizer taken at different e-liquid injection speeds; generate an internal e-liquid filling video of the atomizer at each e-liquid injection speed using a variational autoencoder based on the multiple consecutive X-ray images of the atomizer taken at each e-liquid injection speed; determine the internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds using an e-liquid filling information determination model based on the internal e-liquid filling video of the atomizer at each e-liquid injection speed; determine a target e-liquid filling speed based on the internal e-liquid filling information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid injection speeds; and fill the e-cigarette with e-liquid based on the target e-liquid filling speed.
[0058] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this specification are not intended to limit the order of the processes and methods described herein. Although various examples have been discussed in the foregoing disclosure of some embodiments of the invention that are currently considered useful, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the spirit and scope of the embodiments described herein. For example, while the system components described above can be implemented using hardware devices, they can also be implemented solely using software solutions, such as installing the described system on existing servers or mobile devices.
[0059] Similarly, it should be noted that, in order to simplify the description disclosed herein and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of embodiments in this specification may sometimes combine multiple features into a single embodiment, drawing, or description thereof. However, this method of disclosure does not imply that the subject matter of this specification requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of a single embodiment disclosed above.
[0060] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be illustrative rather than limiting, and should be considered consistent with the teachings of this specification. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.
Claims
1. A method for manufacturing electronic cigarettes based on image analysis, characterized in that, include: Acquire multiple consecutive X-ray images of the atomizer at different e-liquid injection speeds; Based on multiple consecutive X-ray images of the atomizer taken at each e-liquid injection speed, a variational autoencoder is used to generate an internal e-liquid injection video of the atomizer at each e-liquid injection speed. Based on the internal e-liquid filling video of the atomizer at each e-liquid filling speed, the e-liquid filling information determination model is used to determine the internal e-liquid filling information of the atomizer at each e-liquid filling speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid filling speeds; The target e-liquid filling speed is determined based on the internal e-liquid filling information of the atomizer at each e-liquid filling speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid filling speeds. This determination includes: A knowledge graph is constructed, which includes multiple nodes and multiple edges between the nodes. Each node represents an e-liquid injection speed, and the edges between nodes represent the similarity of the internal e-liquid injection information of the atomizer at the e-liquid injection speed. The node features of each node include the internal e-liquid injection information of the atomizer at that e-liquid injection speed. The target oil injection speed is determined by processing the knowledge graph based on a graph autoencoder. The e-cigarette is filled with e-liquid based on the target filling speed.
2. The method for preparing an electronic cigarette based on image analysis as described in claim 1, characterized in that, The oil injection information determination model is a long short-term neural network model.
3. The method for preparing an electronic cigarette based on image analysis as described in claim 1, characterized in that, The internal e-liquid filling information of the atomizer at each e-liquid injection speed includes e-liquid uniformity, bubble formation information, e-liquid filling degree, and e-liquid sealing degree.
4. An electronic cigarette manufacturing system based on image analysis, characterized in that, include: The acquisition module is used to acquire multiple consecutive X-ray images of the atomizer at different e-liquid injection speeds; The generation module is used to generate an internal e-liquid filling video of the atomizer at each e-liquid filling speed using a variational autoencoder based on multiple consecutive X-ray images of the atomizer taken at each e-liquid filling speed. The information determination module is used to determine the internal e-liquid filling information of the atomizer at each e-liquid filling speed and the similarity of the internal e-liquid filling information of the atomizer at different e-liquid filling speeds based on the internal e-liquid filling video of the atomizer at each e-liquid filling speed using the e-liquid filling information determination model. The e-liquid injection speed determination module is used to determine a target e-liquid injection speed based on the internal e-liquid injection information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid injection information of the atomizer at different e-liquid injection speeds. The determination of the target e-liquid injection speed based on the internal e-liquid injection information of the atomizer at each e-liquid injection speed and the similarity of the internal e-liquid injection information of the atomizer at different e-liquid injection speeds includes: A knowledge graph is constructed, which includes multiple nodes and multiple edges between the nodes. Each node represents an e-liquid injection speed, and the edges between nodes represent the similarity of the internal e-liquid injection information of the atomizer at the e-liquid injection speed. The node features of each node include the internal e-liquid injection information of the atomizer at that e-liquid injection speed. The target oil injection speed is determined by processing the knowledge graph based on a graph autoencoder. The e-liquid filling module is used to fill the e-cigarette with e-liquid based on the target e-liquid filling speed.
5. The image analysis-based electronic cigarette manufacturing system as described in claim 4, characterized in that, The oil injection information determination model is a long short-term neural network model.
6. The image analysis-based electronic cigarette manufacturing system as described in claim 4, characterized in that, The internal e-liquid filling information of the atomizer at each e-liquid injection speed includes e-liquid uniformity, bubble formation information, e-liquid filling degree, and e-liquid sealing degree.
7. An electronic device, characterized in that, include: processor; Memory; And a computer program; wherein the computer program is stored in the memory and configured to be executed by the processor to implement the image analysis-based electronic cigarette preparation method as described in any one of claims 1 to 3.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the image analysis-based electronic cigarette preparation method as described in any one of claims 1 to 3.