Fire extinguisher inspection method based on Internet of Things and AI visual inspection
Through the Internet of Things and AI visual inspection technology, identifying whether the fire extinguisher pressure gauge pointer is in the green area, solving the problem of difficult to ensure inspection quality in the existing fire extinguisher inspection mode, realizing intelligent supervision and management closed loop, improving detection accuracy and efficiency.
Patent Information
- Application Number
- CN202411179988.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-27
- Publication Date
- 2024-12-06
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The existing fire extinguisher inspection model lacks intelligent supervision, which makes it difficult to ensure the inspection quality, and the air leakage cannot be detected in time, and there are problems with closed-loop management and trace residues.
Using an Internet of Things and AI visual detection method, we use image data preprocessing and convolutional neural network to identify whether the fire extinguisher pressure gauge pointer is in the green area, realizing intelligent supervision and management traces.
It improves the accuracy of the identification of the fire extinguisher pressure gauge pointer, realizes closed-loop management of inspections, replaces traditional paper inspection cards, and improves inspection efficiency and accuracy.
Smart Images

Figure CN119091209A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of fire extinguisher inspection, and in particular relates to a fire extinguisher inspection method based on the Internet of Things and AI visual detection. Background Art
[0002] At present, the management mode of portable fire extinguishers (GB4351) and trolley-type fire extinguishers (GB8109) is as follows: fire extinguisher administrators regularly go to the site to check the fire extinguishers (mostly once a month). The inspection items mainly include checking whether there are fire extinguishers on site, whether the number of fire extinguishers is sufficient, and whether the fire extinguishers can be used (the key indicator "driving gas pressure", that is, the pointer of the pressure indicator, should be in the green area of 1.0-1.4MPa on the dial; excluding carbon dioxide fire extinguishers), and then record the inspection results on the paper inspection record card attached to the fire extinguisher.
[0003] The manual inspection process is a static management mode, and its advantage is that it can inspect the accessories and appearance of the fire extinguisher on site. However, there are obvious defects: the inspection process is mostly done by one person, and the work quality cannot be guaranteed without supervision. There may be situations such as failure to inspect at the prescribed time, failure to record the actual situation, and poor review of the work quality. The inspection results cannot be synchronized and verticalized through the Internet of Things, and cannot be managed with traces and closed-loop management. There are situations where fire-fighting equipment is not found to have leaked and cannot be replaced in time, and there is a hidden danger that the fire-fighting equipment cannot effectively extinguish the fire. Summary of the invention
[0004] In view of the problems existing in the above-mentioned prior art, the present invention provides a fire extinguisher inspection method based on the Internet of Things and AI visual detection. The technical problem to be solved by the present invention is how to realize intelligent supervision of the pressure condition of the fire extinguisher.
[0005] In order to solve the above technical problems, the present invention provides a fire extinguisher inspection method based on the Internet of Things and AI visual detection, comprising the following steps:
[0006] A fire extinguisher inspection method based on the Internet of Things and AI visual detection, characterized in that it includes the following steps:
[0007] Step S1: collecting a number of image data including a fire extinguisher pressure gauge, and then preprocessing the image data to form a data set, and dividing it into a training set, a test set and a validation set;
[0008] Step S2: extract key features from the training set;
[0009] Step S3: manually mark whether the pointer of the fire extinguisher pressure gauge in each image in the training set is in the green area; input the extracted key features and the marked information into the convolutional neural network CNN to train the model, and obtain the fire extinguisher pressure gauge pointer recognition model;
[0010] Step S4: using the validation set to tune the fire extinguisher pressure gauge pointer recognition model, and improving the recognition accuracy by adjusting the fire extinguisher pressure gauge pointer recognition model parameters and structure;
[0011] Step S5: using the test set to evaluate the performance of the fire extinguisher pressure gauge pointer recognition model, and obtaining a fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements;
[0012] Step S6: The inspector takes an image of the fire extinguisher pressure gauge on site and inputs it into a fire extinguisher pressure gauge pointer recognition model that meets the assessment requirements;
[0013] Step S7: The fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements determines whether the fire extinguisher pressure gauge pointer is in the green area;
[0014] Step S8: Using the images collected on-site by the inspectors, the fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements is updated.
[0015] Furthermore, in step S1, the image data includes the pointer position, and the image data is an image captured under set light conditions.
[0016] Furthermore, in step S1, the preprocessing includes denoising, contrast enhancement and color correction.
[0017] Furthermore, in step S2, the key features include the boundary, scale lines, and the position and shape of the pointer of the pressure gauge.
[0018] Furthermore, in step S2, key features are obtained through edge detection, morphological operations, and color segmentation algorithms.
[0019] Furthermore, the step S2, extracting key features includes the following steps:
[0020] Step S21: determining a joint region including multiple objects in the image to be identified;
[0021] Step S22: extracting visual feature points of the joint area, and converting the visual feature points into a normal distribution of multiple visual feature points corresponding to multiple visual relationships;
[0022] Step S23: extracting semantic feature points of each visual relationship, and converting the semantic feature points into a normal distribution of semantic feature points corresponding to the visual relationship;
[0023] Step S24: determining a normal distribution of semantic feature points that matches the normal distribution of visual feature points among the normal distributions of multiple semantic feature points that correspond one-to-one to the multiple visual relationships;
[0024] Step S25: determining the visual relationship corresponding to the normal distribution of the matched semantic feature points as the visual relationship between the multiple objects.
[0025] Furthermore, in step S5, the performance evaluation indicators of the fire extinguisher pressure gauge pointer recognition model include: accuracy, recall rate, and F1 score.
[0026] Furthermore, in step S6, the inspector scans and photographs the pressure gauge within a specified cone angle to obtain optimal image data.
[0027] Furthermore, in step S6, the inspector uses a PDA device or a mobile phone to collect the pressure gauge image.
[0028] The present invention provides a fire extinguisher inspection method based on the Internet of Things and AI visual detection. By using AI visual technology to establish a fire extinguisher pressure gauge pointer recognition model to distinguish whether the pointer of the fire extinguisher pressure gauge is in the green area, the algorithm is used to determine whether the fire extinguisher is leaking, thereby realizing intelligent supervision of the fire extinguisher and avoiding the problem of difficult quality assurance of manual inspections; and realizing management traces of inspections. Closed-loop management of inspections is realized; traditional paper inspection cards are upgraded to an Internet of Things inspection system platform; pressure gauge images can be uploaded through a dedicated PDA device or a mobile phone, which is simple and convenient; the detection accuracy of visual relationships can be improved by this method. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 The present invention is a flow chart of a fire extinguisher inspection method based on the Internet of Things and AI visual detection. DETAILED DESCRIPTION
[0030] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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.
[0031] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. 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.
[0032] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.
[0033] In the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be directly connected or indirectly connected through an intermediate medium, it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0034] In order to better understand the purpose, structure and function of the present invention, the fire extinguisher inspection method based on the Internet of Things and AI visual detection of the present invention is further described in detail below in conjunction with the accompanying drawings.
[0035] Embodiment 1:
[0036] like Figure 1 As shown, the present invention provides a fire extinguisher inspection method based on the Internet of Things and AI visual detection, comprising the following steps:
[0037] Step S1: collecting a number of image data including a fire extinguisher pressure gauge, and then preprocessing the image data to form a data set, and dividing it into a training set, a test set and a validation set;
[0038] Step S2: extract key features from the training set;
[0039] Step S3: manually mark whether the pointer of the fire extinguisher pressure gauge in each image in the training set is in the green area; input the extracted key features and the marked information into the convolutional neural network CNN to train the model, and obtain the fire extinguisher pressure gauge pointer recognition model;
[0040] Step S4: using the validation set to tune the fire extinguisher pressure gauge pointer recognition model, and improving the recognition accuracy by adjusting the fire extinguisher pressure gauge pointer recognition model parameters and structure;
[0041] Step S5: using the test set to evaluate the performance of the fire extinguisher pressure gauge pointer recognition model, and obtaining a fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements;
[0042] Step S6: The inspector takes an image of the fire extinguisher pressure gauge on site and inputs it into a fire extinguisher pressure gauge pointer recognition model that meets the assessment requirements;
[0043] Step S7: The fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements determines whether the fire extinguisher pressure gauge pointer is in the green area;
[0044] Step S8: Using the images collected on-site by the inspectors, the fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements is updated.
[0045] Embodiment 2:
[0046] like Figure 1 As shown, the present invention provides a fire extinguisher inspection method based on the Internet of Things and AI visual detection, comprising the following steps:
[0047] Step S1: collecting a number of image data including a fire extinguisher pressure gauge, and then preprocessing the image data to form a data set, and dividing it into a training set, a test set and a validation set;
[0048] Step S2: extract key features from the training set;
[0049] Step S3: manually mark whether the pointer of the fire extinguisher pressure gauge in each image in the training set is in the green area; input the extracted key features and the marked information into the convolutional neural network CNN to train the model, and obtain the fire extinguisher pressure gauge pointer recognition model;
[0050] Step S4: using the validation set to tune the fire extinguisher pressure gauge pointer recognition model, and improving the recognition accuracy by adjusting the fire extinguisher pressure gauge pointer recognition model parameters and structure;
[0051] Step S5: using the test set to evaluate the performance of the fire extinguisher pressure gauge pointer recognition model, and obtaining a fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements;
[0052] Step S6: The inspector takes an image of the fire extinguisher pressure gauge on site and inputs it into a fire extinguisher pressure gauge pointer recognition model that meets the assessment requirements;
[0053] Step S7: The fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements determines whether the fire extinguisher pressure gauge pointer is in the green area;
[0054] Step S8: Using the images collected on-site by the inspectors, the fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements is updated.
[0055] The difference between this embodiment and the first embodiment is that:
[0056] In step S1, the image data includes the pointer position, and the image data is an image captured under set light conditions.
[0057] In step S1, the preprocessing includes denoising, contrast enhancement and color correction to improve image quality and reduce the difficulty of subsequent processing;
[0058] In step S2, the key features include the boundary, scale lines, and position and shape of the pointer of the pressure gauge;
[0059] In step S2, key features are obtained by edge detection, morphological operation, color segmentation and other algorithms;
[0060] The step S2, extracting key features comprises the following steps:
[0061] Step S21: determining a joint region including multiple objects in the image to be identified;
[0062] Step S22: extracting visual feature points of the joint area, and converting the visual feature points into a normal distribution of multiple visual feature points corresponding to multiple visual relationships;
[0063] Step S23: extracting semantic feature points of each visual relationship, and converting the semantic feature points into a normal distribution of semantic feature points corresponding to the visual relationship;
[0064] Step S24: determining a normal distribution of semantic feature points that matches the normal distribution of visual feature points among the normal distributions of multiple semantic feature points that correspond one-to-one to the multiple visual relationships;
[0065] Step S25: determining the visual relationship corresponding to the normal distribution of the matched semantic feature points as the visual relationship between the multiple objects.
[0066] In step S5, the performance evaluation indicators of the fire extinguisher pressure gauge pointer recognition model include: accuracy, recall rate, and F1 score.
[0067] In step S6, the inspector scans and photographs the pressure gauge within a specified cone angle to obtain optimal image data.
[0068] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention are included in the protection scope of the present invention.
Claims
1. A fire extinguisher inspection method based on the Internet of Things and AI visual detection, characterized in that: The following steps are involved: Step S1: collecting a number of image data including a fire extinguisher pressure gauge, and then preprocessing the image data to form a data set, and dividing it into a training set, a test set and a validation set; Step S2: extract key features from the training set; Step S3: manually mark whether the pointer of the fire extinguisher pressure gauge in each image in the training set is in the green area; input the extracted key features and the marked information into the convolutional neural network CNN to train the model, and obtain the fire extinguisher pressure gauge pointer recognition model; Step S4: using the validation set to tune the fire extinguisher pressure gauge pointer recognition model, and improving the recognition accuracy by adjusting the fire extinguisher pressure gauge pointer recognition model parameters and structure; Step S5: using the test set to evaluate the performance of the fire extinguisher pressure gauge pointer recognition model, and obtaining a fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements; Step S6: The inspector takes an image of the fire extinguisher pressure gauge on site and inputs it into a fire extinguisher pressure gauge pointer recognition model that meets the assessment requirements; Step S7: The fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements determines whether the fire extinguisher pressure gauge pointer is in the green area; Step S8: Using the images collected on-site by the inspectors, the fire extinguisher pressure gauge pointer recognition model that meets the evaluation requirements is updated.
2. The fire extinguisher inspection method based on the Internet of Things and AI visual detection according to claim 1 is characterized in that: In step S1, the image data includes the pointer position, and the image data is an image captured under set light conditions.
3. The fire extinguisher inspection method based on the Internet of Things and AI visual detection according to claim 2 is characterized in that: In step S1, the preprocessing includes denoising, contrast enhancement and color correction.
4. The fire extinguisher inspection method based on the Internet of Things and AI visual detection according to claim 1 is characterized in that: In step S2, the key features include the boundary, scale lines, and the position and shape of the pointer of the pressure gauge.
5. The fire extinguisher inspection method based on the Internet of Things and AI visual detection according to claim 1 is characterized in that: In step S2, key features are obtained through edge detection, morphological operations, and color segmentation algorithms.
6. The fire extinguisher inspection method based on the Internet of Things and AI visual detection according to claim 1 is characterized in that: The step S2, extracting key features comprises the following steps: Step S21: determining a joint region including multiple objects in the image to be identified; Step S22: extracting visual feature points of the joint area, and converting the visual feature points into a normal distribution of multiple visual feature points corresponding to multiple visual relationships; Step S23: extracting semantic feature points of each visual relationship, and converting the semantic feature points into a normal distribution of semantic feature points corresponding to the visual relationship; Step S24: determining a normal distribution of semantic feature points that matches the normal distribution of visual feature points among the normal distributions of multiple semantic feature points that correspond one-to-one to the multiple visual relationships; Step S25: determining the visual relationship corresponding to the normal distribution of the matched semantic feature points as the visual relationship between the multiple objects.
7. The fire extinguisher inspection method based on the Internet of Things and AI visual detection according to claim 1 is characterized in that: In step S5, the performance evaluation indicators of the fire extinguisher pressure gauge pointer recognition model include: accuracy, recall rate, and F1 score.
8. The fire extinguisher inspection method based on the Internet of Things and AI visual detection according to claim 1 is characterized in that: In step S6, the inspector scans and photographs the pressure gauge within a specified cone angle to obtain optimal image data.
9. The fire extinguisher inspection method based on the Internet of Things and AI visual detection according to claim 1 is characterized in that: In step S6, the inspector uses a PDA device or a mobile phone to collect the pressure gauge image.
