Smoke concentration quantitative analysis system and method based on image recognition
By designing a smoke concentration quantization analysis system based on image recognition, using deep convolutional neural network and smoke concentration model, the problem of difficult and misjudgment of smoke concentration measurement in the existing technology is solved, and accurate and reliable smoke concentration measurement and dynamic change display are achieved.
Patent Information
- Application Number
- CN202411764963.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-04
- Publication Date
- 2025-03-25
AI Technical Summary
In the prior art, smoke concentration measurement based on video images is difficult to achieve in a real environment and is prone to misjudgment, mainly due to the limitations of background and atmospheric light changes, smoke blur images, smoke boundary blur and artificial extraction of features.
Design a smoke concentration quantization analysis system based on image recognition, including image acquisition, image processing, smoke detection, concentration quantization, analysis and display and system management modules. The smoke area is identified through deep convolutional neural network, combined with the smoke concentration model to extract image features, quantify the smoke concentration, and display the concentration change trend and distribution through visualization technology.
It realizes the accuracy and reliability of smoke concentration measurement, can effectively identify and quantify smoke concentration in a real environment, provide intuitive dynamic display of changes, and achieve efficient management and data tracking through cloud storage.
Smart Images
Figure CN119693767A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of smoke concentration recognition, and in particular relates to a smoke concentration quantitative analysis system and method based on image recognition. Background Art
[0002] Smoke density is one of the essential characteristics of smoke and is an important criterion for measuring fire severity and fire detection. Compared with smoke detection, smoke density provides richer and more accurate information for fire simulation and personnel evacuation, but smoke density measurement also requires more data information than simple smoke detection.
[0003] The traditional method of measuring smoke density based on video images mainly detects based on the image similarity or image grayscale changes in the presence and absence of smoke, which requires known atmospheric light and background images. In real environments, the background and atmospheric light of video images often change, so the above method is not suitable for measurement in real environments. Moreover, when directly measuring smoke density based on image data alone, smoke will blur the image and reduce the image contrast. In addition, smoke has no specific appearance shape and the boundary is very fuzzy. The translucent nature of smoke itself will also cause the smoke texture to mix with the background, making direct measurement of smoke density based on video images a difficult problem. The training data in smoke segmentation based on deep learning is mainly synthetic data, which is somewhat different from the real smoke situation. In addition, most of the existing algorithms for measuring smoke density are based on models with artificially extracted features, which are prone to bottlenecks and lack generalization, because artificially extracted image features are not sufficient to represent the complex changes in smoke images.
[0004] Therefore, it is necessary to propose a smoke concentration quantitative analysis system and method based on image recognition to solve the problem that smoke concentration measurement based on video images in the prior art is difficult to achieve and prone to misjudgment.
[0005] The above information disclosed in this background technology is only used to increase the understanding of the background technology of the present invention and therefore, it may include information that does not constitute the prior art known to ordinary technicians in this field. Summary of the invention
[0006] The purpose of the present invention is to provide a smoke concentration quantitative analysis system and method based on image recognition to solve the problems raised in the above background technology.
[0007] To achieve the above object, the present invention provides the following technical solutions:
[0008] A smoke concentration quantitative analysis system based on image recognition, comprising: an image acquisition module, an image processing module, a smoke detection module, a concentration quantification module, an analysis and display module, and a system management module;
[0009] The image acquisition module is used to acquire real-time image or video data through a camera, and send the image or video data to the image processing module;
[0010] The image processing module is used to optimize the input image or video data and send the image or video data to the smoke detection module;
[0011] The smoke detection module is used to detect the smoke area in the received image or video data by using image recognition technology, and send the smoke area to the concentration quantification module;
[0012] The concentration quantification module is used to extract image features based on the detected smoke area using a smoke concentration model to quantify the smoke concentration;
[0013] The analysis and display module is used to process and display the smoke concentration data, and provide visualization information on the concentration change trend and smoke distribution;
[0014] The system management module is used for overall management and monitoring of system operation, timely synchronization of all data, and storage in the cloud.
[0015] Preferably, the image acquisition module is also used to configure the working mode of the camera according to the type of the camera, set the frequency and resolution of image acquisition, and real-time video data acquisition; the integrated monitoring mechanism checks the working status of the camera in real time and handles possible faults or frame loss problems.
[0016] Preferably, the image processing module is also used to use a Gaussian filtering algorithm to remove noise in the image or video data; enhance the image or video data through histogram equalization and edge enhancement technology; and use an image segmentation algorithm to crop the image or video data to remove background debris.
[0017] Preferably, the smoke detection module is also used to identify the smoke area on the image or video data using a deep convolutional neural network; perform real-time processing on each frame of video image, and track smoke dynamics through inter-frame difference detection; mark the smoke area, and pass the extracted image features to the concentration quantification module.
[0018] Preferably, the concentration quantification module is also used to establish a mapping relationship between the smoke concentration and the image features to obtain a quantitative range of smoke concentration values; use the pre-trained smoke concentration model to analyze the smoke area to obtain an estimated value of the smoke concentration; and classify and locate the smoke area in combination with the preset smoke concentration value range.
[0019] Preferably, the analysis and display module is also used to record and store the smoke concentration data, provide historical data query and statistical analysis functions; construct a graphical interface through visualization technology to display the changing trend of the estimated value, real-time image or video data; when the estimated value exceeds a preset threshold, trigger an alarm function and send an alarm message to the user.
[0020] Preferably, the system management module is also used to configure the MySQL database as a buffer, and upload the data in the MySQL database to the cloud according to a preset upload cycle; when a system abnormality is detected, an alarm is automatically triggered, and the system abnormality information is sent to the administrator; when a new version or patch is detected, it is automatically downloaded and pushed to each system module for updating.
[0021] A smoke concentration quantitative analysis method based on image recognition, comprising:
[0022] Step 1: Use a camera to collect image or video data in the environment in real time, and pre-process the image or video data;
[0023] Step 2: Use a deep convolutional neural network to analyze the above image or video data to obtain the smoke area, and analyze the real-time changes of the smoke area through dynamic tracking technology;
[0024] Step 3: Use the trained smoke density model to analyze the smoke area, calculate the smoke density estimate, and locate the concentration range in which it is located;
[0025] Step 4: Use visualization technology to display the change trend of the estimated value of smoke concentration on a graphical interface, and send an alarm message to the user when it reaches a preset threshold;
[0026] Step 5. Use the MySQL database as a data buffer, upload the stored data to the cloud regularly, monitor the system operation status in real time, and automatically trigger an alarm when an abnormality is found.
[0027] Preferably, in the above method, an image segmentation algorithm is applied to remove background debris in the image and retain the area related to the smoke; the smoke area in the video is tracked in real time through inter-frame difference detection to identify the dynamic changes of the smoke in the video; the identified smoke area is marked and the image features of the area are extracted; the cloud is connected through an API interface and the uploaded sensitive data is encrypted.
[0028] Compared with the prior art, the present invention has the following beneficial effects:
[0029] The present invention ensures the clarity and accuracy of input data through real-time image or video data collection and optimization processing, and adopts advanced image recognition technology and smoke concentration model to make the measurement of smoke concentration more accurate and highly reliable; through real-time analysis and display of the changing trend and distribution of smoke concentration, users can intuitively and comprehensively grasp the dynamic changes of smoke concentration; through overall integration and cloud data storage, not only the efficient management of the smoke monitoring process is achieved, but also the continuous tracking and remote access of data are ensured.
[0030] The above summary is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments and features described above, further aspects, embodiments and features of the present invention will be readily apparent by reference to the accompanying drawings and the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] Figure 1 This is a framework diagram of the smoke concentration quantitative analysis system based on image recognition of the present invention;
[0032] Figure 2 This is a flow chart of the smoke concentration quantitative analysis method based on image recognition of the present invention. DETAILED DESCRIPTION
[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0034] Embodiment 1:
[0035] See also Figure 1 As shown, a smoke concentration quantitative analysis system based on image recognition includes: an image acquisition module, an image processing module, a smoke detection module, a concentration quantification module, an analysis and display module, and a system management module;
[0036] The image acquisition module is used to obtain real-time image or video data through a camera and send the image or video data to the image processing module;
[0037] The image processing module is used to optimize the input image or video data and send the image or video data to the smoke detection module;
[0038] The smoke detection module is used to detect the smoke area in the received image or video data through image recognition technology, and send the smoke area to the concentration quantification module;
[0039] The concentration quantification module is used to extract image features based on the detected smoke area using the smoke concentration model and quantify the smoke concentration;
[0040] The analysis and display module is used to process and display the above smoke concentration data, and provide visualization information on concentration change trends and smoke distribution;
[0041] The system management module is used for overall management and monitoring of system operation, timely synchronization of all data, and storage in the cloud.
[0042] The image acquisition module is also used to configure the working mode of the camera according to its type, set the frequency and resolution of image acquisition, and real-time video data acquisition; select the appropriate camera according to the actual situation, and commonly use high-definition cameras, which need to support image or video acquisition under different environmental conditions, such as low light, high contrast, etc. The integrated monitoring mechanism checks the working status of the camera in real time and handles possible faults or frame loss problems.
[0043] The image processing module is also used to remove noise from image or video data using a Gaussian filtering algorithm; enhance image or video data through histogram equalization and edge enhancement technology; crop image or video data using an image segmentation algorithm to remove background debris; and adjust image brightness and contrast according to ambient light conditions.
[0044] The smoke detection module is also used to identify smoke areas in image or video data using a deep convolutional neural network;
[0045] Each frame of video image is processed in real time, and smoke dynamics are tracked through frame difference detection; smoke areas are marked, and the extracted image features are passed to the concentration quantification module.
[0046] The concentration quantification module is also used to establish a mapping relationship between smoke concentration and image features to obtain a quantitative range of smoke concentration values; to analyze the smoke area using a pre-trained smoke concentration model to obtain an estimated value of the smoke concentration; and to classify and locate the smoke area based on the preset smoke concentration value range.
[0047] The smoke density model extracts the global feature map and detail feature map of the smoke area respectively, and fuses the two. The fused feature image is obtained through a convolution layer with a 1×1 convolution kernel, and then the dimensionality reduction is performed on it, and finally the estimated value of the smoke density is output.
[0048] Extract global feature maps: The convolution layer consists of a group of 3×3 convolution kernels, the initial number of channels is 32, the step size is 1×1, and ReLu is selected as the activation function. Starting from the second convolution layer, the number of channels increases layer by layer. The first pooling layer is configured between the second and third convolution layers. Similarly, the second pooling layer is configured between the fourth and fifth convolution layers, and the third pooling layer is configured after the seventh convolution layer.
[0049] Extract detail feature maps: The convolution kernel size of the first convolution layer is set to 5. The subsequent four convolution layers are composed of a group of 3×3 convolution kernels. The output of the fifth convolution layer and the output of the first convolution layer are used together as the input of the sixth convolution layer. The convolution kernel of the sixth convolution layer is 1×1, and the number of channels is the same as that of the fifth convolution layer. The first pooling layer is configured for pooling, and the convolution kernel is 3 for the seventh and eighth convolution layers. The second pooling layer is configured for pooling, and the convolution kernel is 5 for the ninth convolution layer. The third pooling layer is configured for pooling.
[0050] The smoke concentration value ranges from 0 to 1, 0 means there is no smoke in this area, and 1 means the smoke concentration in this area reaches the maximum value. The smoke concentration value range is divided into: 0-0.15 is defined as a smoke-free area, 0.15-0.55 is defined as a low-concentration smoke area, 0.55-0.85 is defined as a high-concentration smoke area, and 0.85-1 is defined as an extremely high-concentration smoke area.
[0051] The analysis and display module is also used to record and store smoke concentration data, provide historical data query and statistical analysis functions; build a graphical interface through visualization technology to display the changing trend of the estimated value, real-time image or video data; when the estimated value exceeds the preset threshold, the alarm function is triggered and an alarm message is sent to the user.
[0052] The system management module is also used to configure the MySQL database as a buffer and upload the data in the MySQL database to the cloud according to the preset upload cycle; when a system abnormality is detected, an alarm is automatically triggered and the system abnormality information is sent to the administrator; when a new version or patch is detected, it is automatically downloaded and pushed to each system module for updating.
[0053] Embodiment 2:
[0054] See also Figure 2 As shown, a smoke concentration quantitative analysis method based on image recognition includes:
[0055] Step 1: Use a camera to collect image or video data in the environment in real time and pre-process the image or video data;
[0056] Step 2: Use a deep convolutional neural network to analyze the above image or video data to obtain the smoke area, and analyze the real-time changes of the smoke area through dynamic tracking technology;
[0057] Step 3: Use the trained smoke density model to analyze the smoke area, calculate the estimated smoke density, and locate the concentration range;
[0058] Step 4: Use visualization technology to display the change trend of the estimated value of smoke concentration on a graphical interface, and send an alarm message to the user when it reaches a preset threshold;
[0059] Step 5. Use the MySQL database as a data buffer, upload the stored data to the cloud regularly, monitor the system operation status in real time, and automatically trigger an alarm when an abnormality is found.
[0060] In the above method, an image segmentation algorithm is applied to remove background debris in the image and retain the area related to the smoke; the smoke area in the video is tracked in real time through inter-frame difference detection to identify the dynamic changes of the smoke in the video; the identified smoke area is marked and the image features of the area are extracted; the API interface is used to connect to the cloud and the uploaded sensitive data is encrypted.
[0061] As can be seen from the above, the present invention ensures the clarity and accuracy of input data through real-time image or video data collection and optimization processing, and adopts advanced image recognition technology and smoke concentration model to make the measurement of smoke concentration more accurate and highly reliable; through real-time analysis and display of the changing trend and distribution of smoke concentration, users can intuitively and comprehensively grasp the dynamic changes of smoke concentration; through overall integration and cloud data storage, not only the efficient management of the smoke monitoring process is achieved, but also the continuous tracking and remote access of data are ensured.
[0062] Embodiment 3:
[0063] The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, each process of the above-mentioned quantitative analysis system embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it is not repeated here. Among them, the computer-readable storage medium, such as a read-only memory (Read-Only Memory, referred to as ROM), a random access memory (Random Access Memory, referred to as RAM), a magnetic disk or an optical disk, etc.
[0064] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "examples", "specific examples" or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine different embodiments or examples described in this specification and the features of different embodiments or examples, unless they are contradictory.
[0065] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems) and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a program, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A way to specify functions in one or more boxes.
[0066] In the drawings of the embodiments disclosed in the present invention, only the structures involved in the embodiments disclosed in the present invention are involved, and other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of the present invention can be combined with each other.
[0067] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.
[0068] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A smoke concentration quantitative analysis system based on image recognition, characterized in that: include: Image acquisition module, image processing module, smoke detection module, concentration quantification module, analysis and display module, system management module; The image acquisition module is used to acquire real-time image or video data through a camera, and send the image or video data to the image processing module; The image processing module is used to optimize the input image or video data and send the image or video data to the smoke detection module; The smoke detection module is used to detect the smoke area in the received image or video data by using image recognition technology, and send the smoke area to the concentration quantification module; The concentration quantification module is used to extract image features based on the detected smoke area using a smoke concentration model to quantify the smoke concentration; The analysis and display module is used to process and display the smoke concentration data, and provide visualization information on the concentration change trend and smoke distribution; The system management module is used for overall management and monitoring of system operation, timely synchronization of all data, and storage in the cloud.
2. The smoke concentration quantitative analysis system based on image recognition according to claim 1 is characterized in that: The image acquisition module is also used for: Configuring the working mode of the camera according to its type, setting the frequency and resolution of image acquisition, and real-time video data acquisition; The integrated monitoring mechanism checks the working status of the camera in real time and handles possible failures or frame loss problems.
3. The smoke concentration quantitative analysis system based on image recognition according to claim 2 is characterized in that: The image processing module is also used for: Using a Gaussian filtering algorithm to remove noise from the image or video data; Performing enhancement processing on the image or video data by using histogram equalization and edge enhancement technology; The image or video data is cropped using an image segmentation algorithm to remove background clutter.
4. The smoke concentration quantitative analysis system based on image recognition according to claim 3 is characterized in that: The smoke detection module is also used for: Using a deep convolutional neural network to identify the smoke area on the image or video data; Perform real-time processing on each frame of video image and track smoke dynamics through inter-frame difference detection; The smoke area is marked, and the extracted image features are passed to the concentration quantification module.
5. The smoke concentration quantitative analysis system based on image recognition according to claim 4 is characterized in that: The concentration quantification module is also used for: Establishing a mapping relationship between the smoke concentration and the image feature to obtain a quantitative range of smoke concentration values; Analyze the smoke area using the pre-trained smoke density model to obtain an estimated value of the smoke density; The smoke area is classified and located in combination with the preset smoke concentration value range.
6. The smoke concentration quantitative analysis system based on image recognition according to claim 5 is characterized in that: The analysis and display module is also used for: Record and store the smoke concentration data, and provide historical data query and statistical analysis functions; Constructing a graphical interface through visualization technology to display the changing trend of the estimated value, real-time image or video data; When the estimated value exceeds a preset threshold, an alarm function is triggered and an alarm message is sent to the user.
7. The smoke concentration quantitative analysis system based on image recognition according to claim 6 is characterized in that: The system management module is also used for: Configuring a MySQL database as a buffer, and uploading the data in the MySQL database to the cloud according to a preset upload cycle; When a system anomaly is detected, an alarm is automatically triggered and the system anomaly information is sent to the administrator; When a new version or patch is detected, it is automatically downloaded and pushed to each system module for updating.
8. A method for quantitative analysis of smoke concentration based on image recognition, characterized in that: include: Step 1: Use a camera to collect image or video data in the environment in real time, and pre-process the image or video data; Step 2: Use a deep convolutional neural network to analyze the above image or video data to obtain the smoke area, and analyze the real-time changes of the smoke area through dynamic tracking technology; Step 3: Use the trained smoke density model to analyze the smoke area, calculate the smoke density estimate, and locate the concentration range in which it is located; Step 4: Use visualization technology to display the change trend of the estimated value of smoke concentration on a graphical interface, and send an alarm message to the user when it reaches a preset threshold; Step 5. Use the MySQL database as a data buffer, upload the stored data to the cloud regularly, monitor the system operation status in real time, and automatically trigger an alarm when an abnormality is found.
9. The method for quantitative analysis of smoke concentration based on image recognition according to claim 8, characterized in that: In the above method: Apply image segmentation algorithm to remove background debris in the image and retain the area related to smoke; The smoke area in the video is tracked in real time through inter-frame difference detection to identify the dynamic changes of the smoke in the video.
10. The method for quantitative analysis of smoke concentration based on image recognition according to claim 9, characterized in that: In the above method: Marking the identified smoke area and extracting image features of the area; Connect with the cloud through the API interface and encrypt the uploaded sensitive data.