Modularized laboratory intelligent monitoring platform based on deep learning
Through the combination of modular design and deep learning technology, automatic analysis and target recognition of laboratory monitoring systems are achieved, which solves the shortcomings of traditional monitoring systems in laboratory environments, improves monitoring accuracy and efficiency, and improves the safety and management level of laboratory.
Patent Information
- Application Number
- CN202510532704.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-25
- Publication Date
- 2025-07-18
AI Technical Summary
Traditional monitoring systems have problems such as high artificial dependence, insufficient data processing capabilities, slow response speed and insufficient versatility in laboratory environments, which are difficult to meet the special needs of laboratories for safety, equipment maintenance and experimental process normativeness.
The modularly designed platform server module, recognition algorithm module, camera module and WEB module are adopted, combined with deep learning technology, to realize automatic analysis and target recognition of monitoring images, and efficient processing and analysis are carried out through modular design and API interface.
It improves the accuracy and efficiency of monitoring, meets the intelligent management needs of the laboratory, improves the safety and management level of the laboratory, has flexibility and high compatibility, and adapts to the customized needs of different laboratories.
Smart Images
Figure CN120343296A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of intelligent monitoring, and particularly to a modular intelligent monitoring platform for laboratories based on deep learning. Background Art
[0002] With the rapid development of information technology, the importance of intelligent video monitoring systems in the security field has become increasingly prominent. Traditional monitoring systems mainly rely on manual monitoring and have many obvious defects. Since long-term monitoring work is prone to causing operator fatigue and distraction of attention, the monitoring efficiency and accuracy are greatly limited, making it difficult to effectively guarantee the security of the monitored area. At the same time, in the face of a large amount of video data, traditional systems can often only perform basic motion detection, lacking the ability of in-depth analysis of video content and behavior recognition, and unable to detect potential security hazards and abnormal situations in a timely manner. Moreover, traditional monitoring systems have a slow response speed and usually can only record after an event occurs, unable to achieve real-time monitoring and rapid warning, and it is difficult to meet the requirements of modern security management for timeliness and initiative. In addition, traditional monitoring systems lack flexibility and customizability, and it is difficult to adapt to the special monitoring requirements of different laboratories in terms of security, equipment maintenance, and standardization of experimental processes, and cannot provide personalized monitoring services for laboratories.
[0003] In a laboratory environment, there are extremely high requirements for security, equipment maintenance, standardization of experimental processes, and scientific research efficiency. The equipment in the laboratory is often of high value and complex in structure. Any abnormal situation, such as fire, leakage, etc., may cause serious damage to personnel and equipment. Therefore, it is necessary to monitor and maintain the equipment in real time to avoid equipment damage or experimental failure. At the same time, the standardization and traceability of the experimental process are crucial in scientific research activities. Through an effective monitoring system, the operations of experimental personnel can be recorded and supervised to ensure the standardization of the experimental process and the reliability of the results. However, the existing traditional monitoring systems are difficult to meet these complex and strict requirements, and there is an urgent need for a more intelligent and efficient intelligent monitoring platform. Summary of the Invention
[0004] The purpose of the present invention is to provide a modular intelligent monitoring platform for laboratories based on deep learning, aiming to improve the problems of high manual dependence, insufficient data processing ability, slow response speed, and lack of generality existing in traditional monitoring systems.
[0005] The present invention is implemented as follows: A modular laboratory intelligent monitoring platform based on deep learning, including a platform server module, an identification algorithm module, a camera module, and a WEB module with modular design; the camera module is used to collect video data and transmit it to a streaming media server; the WEB module is used for users to initiate requests and display the identification results received by the platform server module; the platform server module is used to process the camera identification requests initiated by users through the WEB page of the WEB module, start the camera module to collect video, and transmit the video to the identification algorithm module through the streaming media server; the identification algorithm module is used to perform real-time automatic detection and identification on the received video, and return the identification results to the platform server module; the platform realizes the efficient processing and analysis of a large amount of video data through modular design and API-based interfaces.
[0006] Further, the camera module serves as the video data acquisition entry, and its working steps are as follows:
[0007] S1. When starting, listen for requests from the platform server, and execute the opening or closing operation of the camera according to the requests;
[0008] S2. After parsing the request information, collect and process the video stream;
[0009] S3. Push the processed video stream to the streaming media server to achieve real-time video acquisition.
[0010] Further, when the camera module starts, it creates an HTTP server. After receiving the request to open the camera, it uses the thread library to start a new thread to be responsible for collecting and processing the video stream, and at the same time the main thread returns a response of successful operation to the platform server.
[0011] Further, the identification algorithm module uses the YOLOv5 algorithm to perform real-time automatic detection and identification on the received video. The specific working steps of the identification algorithm module are as follows:
[0012] S10. When starting, check and load the model file, and then enter the listening state;
[0013] S20. Receive the identification request, perform identification processing on the video frame according to the request, and process and annotate the identification result image;
[0014] S30. Push the annotated identification result video to the streaming media server.
[0015] Further, when the identification algorithm module starts, it checks and loads the identification model file of the YOLOv5 algorithm in the same-level directory, and uses the io library of Python to load the identification model file into the computer memory; it starts an HTTP server through the Flask framework to listen for and receive identification requests from the platform server.
[0016] Further, when the recognition algorithm module executes a recognition task, each recognition task is executed in a new thread, enabling the recognition algorithm module to process multiple concurrent recognition tasks simultaneously.
[0017] Further, the platform server module includes a user management unit and a recognition result processing unit. The user management unit is used to provide login and registration functions, manage user permissions, and allocate monitoring tasks. When starting up, it realizes preprocessing operations before and after function execution by loading a self-developed class aspect component, including user permission verification and login status check. The recognition result processing unit is used to process the video recognition results from the recognition algorithm module, integrate these video recognition results into the WEB interface of the WEB module, enable users to view the monitoring videos and recognition analysis data in real time, and at the same time, realize local storage and automatic expiration management of data by loading a self-developed data cache component.
[0018] Further, the recognition algorithm module and the camera module are developed using Python, and the platform server module uses the Flask framework to implement cross-language service calls to seamlessly integrate and use the recognition algorithm module and the camera module.
[0019] Further, when the platform service module starts up, it loads a self-developed data cache component and a class aspect component to realize local storage of data, automatic expiration, and preprocessing operations of function execution.
[0020] Further, the WEB module has a system setting function unit and a graphical interface presentation unit. The system setting function unit enables users to flexibly adjust monitoring parameters according to the specific requirements and actual situations of different laboratories. The monitoring parameters include the monitoring time period, the angles and resolutions of cameras, and specific rules for target recognition. The graphical interface presentation unit uses data visualization technology to present the results of the recognition algorithm module's processing of the monitoring video to users in an intuitive and clear graphical interface form, enabling users to view and analyze monitoring data more conveniently.
[0021] Compared with the prior art, the beneficial effects of the present invention are as follows:
[0022] 1. The present invention combines deep learning technology with a video monitoring system to realize automatic analysis and target recognition of monitoring images, improves the accuracy and efficiency of monitoring, can meet the urgent needs of laboratory intelligent management, promotes the development and application of intelligent video monitoring technology in laboratory scenarios, and improves the safety and management level of laboratories.
[0023] 2. The present invention uses modular deployment and can be flexibly deployed in various locations of the laboratory. This modular design enables the platform to have high adaptability and scalability, and can be customized for installation according to the specific requirements of different laboratories, thereby maximizing the monitoring efficiency.
[0024] 3. The present invention can realize remote monitoring and management functions through the cloud platform. Managers can view the real-time status of the laboratory at any time and place through mobile devices or computers, and perform remote control and data analysis. This convenient remote monitoring method greatly facilitates the management and maintenance of the laboratory.
[0025] 4. The present invention has high compatibility and can seamlessly dock with hardware devices of different brands and models, such as cameras, sensors, etc. This high compatibility enables the system to adapt to various laboratory environments and reduces the costs of equipment replacement and system upgrade.
[0026] 5. The platform system architecture design of the present invention has high scalability. Users can flexibly add new functional modules according to needs, such as environmental monitoring, energy consumption management, etc., to meet the requirements at different stages. This scalability ensures the long-term applicability and investment value of the platform system. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Figure 1 is the electrical control structure block diagram of the modular laboratory intelligent monitoring platform based on deep learning provided in the embodiment of the present invention;
[0028] Figure 2 is the schematic diagram of the implementation of each function of the modular laboratory intelligent monitoring platform based on deep learning provided in the embodiment of the present invention;
[0029] Figure 3 is the schematic diagram of the process of the camera module of the modular laboratory intelligent monitoring platform based on deep learning provided in the embodiment of the present invention;
[0030] Figure 4 is the schematic diagram of the process of the recognition model module of the modular laboratory intelligent monitoring platform based on deep learning provided in the embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0031] In the present invention, unless otherwise clearly defined and limited, terms such as "installation", "connection", "connection", "fixation" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or integrated; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the internal communication of two components or the interaction relationship between two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0032] The following is a further description in conjunction with the accompanying drawings and specific embodiments:
[0033] As Figure 1 - Figure 2 shown, a modular laboratory intelligent monitoring platform based on deep learning includes a platform server module, an identification algorithm module, a camera module, and a WEB module that adopt a modular design. Through modular design and API-based interfaces, the platform realizes the efficient processing and analysis of a large amount of video data. Figure 2 The identification module in
[0034] is the identification algorithm module, and camera A, camera B, and camera C all belong to the camera module. The modular design has the following characteristics and significance:
[0035] First, it improves system stability and maintainability: The functions of each module are independent. When a certain module fails, only this module needs to be checked and repaired, without affecting the normal operation of other modules. For example, if camera A has a hardware failure, only this camera in the camera module needs to be repaired or replaced, and the platform server module, the identification algorithm module, and the WEB module can still work normally, greatly reducing the risk of overall system failure. At the same time, it is also convenient for technicians to quickly locate problems and shorten the maintenance time.
[0036] Second, it enhances system scalability and flexibility: As the monitoring requirements of the laboratory change or technology is updated, a single module can be easily upgraded or expanded. For example, if it is necessary to improve the accuracy of target recognition in the future, the identification algorithm module can be directly optimized or replaced with a more advanced algorithm version; if a new monitoring area is added to the laboratory, only camera devices need to be added to the camera module and simple configuration and docking are required to incorporate the new monitoring data into system management, without the need for large-scale changes to the entire platform architecture, enabling the platform to flexibly adapt to the diverse needs of different laboratories and future development changes.
[0037] Fourth, it realizes the precise allocation and optimization of resources: Each module can configure corresponding hardware and software resources according to its own functional requirements, avoiding resource waste. For the recognition algorithm module with high computing requirements, a high-performance GPU can be equipped for it to accelerate the operation of the deep learning model; while the camera module can select camera devices with appropriate resolution and frame rate according to the requirements of video acquisition in the monitoring scenario, so as to achieve the optimal utilization of resources and reduce the system construction and operation costs on the premise of ensuring system performance.
[0038] Fifth, it is convenient for system standardization and reuse: The modular design makes each module have clear functional boundaries and interface specifications, which provides a basis for the standardized construction of the system. Modules with the same function can be reused in different laboratory intelligent monitoring projects, not only reducing the workload of repeated development, but also improving the universality and consistency of the system. For example, as long as the recognition algorithm module meets the unified interface standard, it can be applied to multiple laboratory monitoring platforms with different scales and requirements, accelerating the popularization and application of intelligent monitoring technology.
[0039] In the modular laboratory intelligent monitoring platform based on deep learning, the camera module is used to collect video data and transmit it to the streaming media server. The present invention adopts Figure 3 the shown process to realize the mobilization of the camera module. As the video data acquisition entrance, the working steps of the camera module are as follows: S1. Listen to the platform server request at startup and execute the opening or closing operation of the camera according to the request; S2. After parsing the request information, collect and process the video stream; S3. Push the processed video stream to the streaming media server to realize real-time video acquisition. When the camera module starts, it creates an HTTP server. After receiving the request to open the camera, it uses the thread library to start a new thread to be responsible for collecting and processing the video stream, and at the same time the main thread returns a successful operation response to the platform server. In this way, the camera module not only realizes the real-time acquisition of video, but also ensures the smooth transmission of the video stream and the continuity of subsequent processing.
[0040] In the modular laboratory intelligent monitoring platform based on deep learning, the WEB module serves as the user interaction interface of the intelligent monitoring platform. It is responsible for providing real-time video monitoring functions and allowing users to smoothly view and analyze the video stream captured by the camera module and transmitted by the streaming media server on the Web page. At the same time, the WEB module can be used for users to initiate requests and display the recognition results received by the platform server module. The WEB module has a system setting function unit and a graphical interface presentation unit. The system setting function unit enables users to flexibly adjust monitoring parameters according to the specific needs and actual situations of different laboratories. The monitoring parameters include the monitoring time period, the angle and resolution of the camera, and specific rules for target recognition. The graphical interface presentation unit uses data visualization technology to present the results of the recognition algorithm module's processing of the monitored video to users in an intuitive and clear graphical interface form, enabling users to more conveniently view and analyze monitoring data.
[0041] In the modular laboratory intelligent monitoring platform based on deep learning, the platform server module is responsible for the dual responsibilities of communication coordination between modules and user interaction. It is used to process the camera recognition requests initiated by users through the WEB page of the WEB module, start the camera module to collect videos, and transmit the videos to the recognition algorithm module through the streaming media server.
[0042] The recognition algorithm module is used to perform real-time automatic detection and recognition on the received videos and return the recognition results to the platform server module. The present invention adopts Figure 4 the process shown to implement the mobilization of the recognition algorithm module. The recognition algorithm module uses the YOLOv5 algorithm to perform real-time automatic detection and recognition on the received videos. The specific working steps of the recognition algorithm module are as follows: S10. Check and load the model file at startup, and then enter the listening state; S20. Receive the recognition request, perform recognition processing on the video frames according to the request, and process and annotate the recognition result images; S30. Push the annotated recognition result video to the streaming media server. When the recognition algorithm module starts up, it checks and loads the recognition model file of the YOLOv5 algorithm in the same-level directory, and uses the io library of Python to load the recognition model file into the computer memory; starts the HTTP server through the Flask framework, listens for and receives recognition requests from the platform server. When the recognition algorithm module executes recognition tasks, each recognition task is executed in a new thread, so that the recognition algorithm module can process multiple concurrent recognition tasks simultaneously.
[0043] The platform server module includes a user management unit and an identification result processing unit. The user management unit is used to provide login and registration functions, manage user permissions and allocate monitoring tasks. When starting up, it realizes preprocessing operations before and after function execution by loading self-developed class aspect components, including user permission verification and login status check. The identification result processing unit is used to process video identification results from the identification algorithm module, integrate these video identification results into the WEB interface of the WEB module, enable users to view monitoring videos and identification analysis data in real time, and at the same time realize local storage and automatic expiration management of data by loading self-developed data cache components.
[0044] In this embodiment, the identification algorithm module and the camera module are developed using Python. The platform server module uses the Flask framework to achieve cross-language service calls to seamlessly integrate the identification algorithm module and the camera module. When the platform service module starts up, it loads self-developed data cache components and class aspect components to realize local storage of data, automatic expiration, and preprocessing operations of function execution, such as user permission verification and login status check. The platform server module serves as an interface for users to interact with the system, provides functions such as login, registration, and system settings, and at the same time manages the allocation of user permissions and monitoring tasks. In addition, the platform service module is also responsible for processing video identification results from the identification module and integrating these results into the Web interface, enabling users to view monitoring videos and identification analysis data in real time.
[0045] The overall operation of this platform: Flask provides support for the platform system. In the platform system, in addition to completing their respective functions, the camera module and the identification algorithm module need to transmit video information to the video media stream server through video streaming for processing and then realize the display of monitoring videos through video pulling; the platform service module and the WEB module will provide various services such as login and email externally.
[0046] The successful implementation of the design of the present invention depends on the close cooperation between hardware and software. In terms of hardware, the project mainly relies on an advanced camera module, which is responsible for capturing and transmitting video data in real time, providing a basis for the intelligent analysis of the system. The software part is the core to realize the intelligent functions of the project, covering the identification algorithm module based on the YOLOv5 algorithm, cross-language service calls implemented using the Flask framework, and the platform server module built with the Java language. These software components not only realize the effective control and management of hardware, but also provide advanced functions such as real-time video monitoring, system settings, and data visualization through modular design and user-friendly Web interfaces.
[0047] The present invention has the following advantages:
[0048] First, it is modular deployment: It can be flexibly deployed in various locations of the laboratory. This modular design makes the platform highly adaptable and scalable, and it can be customized for installation according to the specific needs of different laboratories, thus maximizing the monitoring efficiency.
[0049] Second, it can achieve remote monitoring and management: The present invention can realize remote monitoring and management functions through the cloud platform. Managers can view the real-time status of the laboratory at any time and place through mobile devices or computers, and conduct remote control and data analysis. This convenient remote monitoring method greatly facilitates the management and maintenance of the laboratory.
[0050] Third, it has high compatibility: The present invention has high compatibility and can seamlessly connect to hardware devices of different brands and models, such as cameras, sensors, etc. This high compatibility enables the system to adapt to various laboratory environments and reduces the costs of equipment replacement and system upgrade.
[0051] Fourth, it has strong scalability: The platform system architecture design has high scalability. Users can flexibly add new functional modules according to needs, such as environmental monitoring, energy consumption management, etc., to meet the needs at different stages. This scalability ensures the long-term applicability and investment value of the platform system.
[0052] In summary, the present invention combines deep learning technology with a video monitoring system to achieve automatic analysis and target recognition of monitoring images, improves the accuracy and efficiency of monitoring, can meet the urgent needs of laboratory intelligent management, promotes the development and application of intelligent video monitoring technology in laboratory scenarios, and enhances the safety and management level of the laboratory.
[0053] The above is only the preferred implementation manner of the present invention and is not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.
Claims
1. A modular intelligent monitoring platform for laboratories based on deep learning, characterized in that, It includes a platform server module, an identification algorithm module, a camera module, and a WEB module that adopt a modular design; the camera module is used to collect video data and transmit it to a streaming media server; the WEB module is used for users to initiate requests and display the identification results received by the platform server module. The platform server module is used to process the camera identification requests initiated by users through the WEB page of the WEB module, start the camera module to collect video, and pass the video to the identification algorithm module through the streaming media server; the identification algorithm module is used to perform real-time automatic detection and identification on the received video, and return the identification results to the platform server module. The platform realizes the efficient processing and analysis of a large amount of video data through modular design and API-based interfaces.
2. The modular laboratory intelligent monitoring platform based on deep learning according to claim 1, characterized in that, The camera module, as the video data acquisition entry, works as follows: S1. Listen for requests from the platform server during startup, and perform the operation of turning on or off the camera according to the requests. S2. After parsing the request information, collect and process the video stream. S3. Push the processed video stream to the streaming media server to achieve real-time video acquisition.
3. The modular laboratory intelligent monitoring platform based on deep learning according to claim 2, wherein, When the camera module starts, it creates an HTTP server. After receiving the request to open the camera, it uses the thread library to start a new thread to be responsible for collecting and processing the video stream, and at the same time the main thread returns a response indicating successful operation to the platform server.
4. A modular laboratory intelligent monitoring platform based on deep learning according to claim 1, characterized in that, The identification algorithm module uses the YOLOv5 algorithm to perform real-time automatic detection and identification on the received video. The specific working steps of the identification algorithm module are as follows: S10. Check and load the model file during startup, and then enter the listening state. S20. Receive the identification request, perform identification processing on the video frame according to the request, and process and annotate the identification result image. S30. Push the annotated identification result video stream to the streaming media server.
5. The modular laboratory intelligent monitoring platform based on deep learning according to claim 4, characterized in that, When the identification algorithm module starts, it checks and loads the identification model file of the YOLOv5 algorithm in the same-level directory, and uses the io library of Python to load the identification model file into the computer memory; it starts an HTTP server through the Flask framework, listens for and receives the identification requests from the platform server.
6. The modular laboratory intelligent monitoring platform based on deep learning according to claim 5, characterized in that, When the identification algorithm module executes the identification task, each identification task is executed in a new thread, so that the identification algorithm module can process multiple concurrent identification tasks simultaneously.
7. The intelligent monitoring platform for a modular laboratory based on deep learning according to claim 1, characterized in that, The platform server module includes a user management unit and an identification result processing unit. The user management unit is used to provide login and registration functions, manage user permissions and the allocation of monitoring tasks, and perform preprocessing operations before and after function execution through loading self-developed class aspect components during startup, including user permission verification and login status check; the identification result processing unit is used to process the video identification results from the identification algorithm module, integrate these video identification results into the WEB interface of the WEB module, enable users to view the monitoring video and identification analysis data in real time, and at the same time realize local storage and automatic expiration management of data through loading self-developed data cache components.
8. The intelligent monitoring platform for modular laboratories based on deep learning according to claim 1, characterized in that, The recognition algorithm module and the camera module are developed using Python. The platform server module utilizes the Flask framework to implement cross-language service calls for seamless integration of the recognition algorithm module and the camera module.
9. The modular laboratory intelligent monitoring platform based on deep learning according to claim 1, characterized in that, When the platform service module starts, it loads the self-developed data cache component and class aspect component to achieve local storage of data, automatic expiration, and preprocessing operations for function execution.
10. A modular laboratory intelligent monitoring platform based on deep learning according to claim 1, characterized in that, The WEB module has a system setting function unit and a graphical interface presentation unit. The system setting function unit enables users to flexibly adjust monitoring parameters according to the specific requirements and actual situations of different laboratories. The monitoring parameters include the monitored time period, the angles and resolutions of the cameras, and specific rules for target recognition. The graphical interface presentation unit uses data visualization technology to present the results of the recognition algorithm module's processing of the monitored video to users in an intuitive and clear graphical interface form, enabling users to view and analyze monitoring data more conveniently.