Coal mine flame pseudo fire interference suppression method based on time sequence consistency analysis

Abstract

The invention relates to a coal mine flame pseudo fire interference suppression method based on time sequence consistency analysis, and belongs to the technical field of coal mine safety monitoring. According to the method, aiming at the problem of high false alarm rate of a flame detection system caused by false fire source interference such as electric welding sparks and mechanical friction in a mine environment, the flame identification accuracy is improved through time sequence consistency analysis. According to the technical scheme, the method comprises the following steps: acquiring an infrared spectrum flame saliency probability graph of continuous image frames in an underground coal mine; extracting a flame candidate region by adopting Otsu adaptive threshold segmentation in combination with morphological optimization and hole filling; and performing time sequence consistency analysis on the continuous frame candidate regions, and calculating the spatial overlapping degree, the morphological stability and the infrared thermal intensity change trend of adjacent frames. According to the method, the false fire false alarm rate is remarkably reduced, the robustness to complex backgrounds and dynamic illumination is enhanced, the method can be independently integrated to an existing infrared flame detection system or used in cooperation with a deep learning model, and the reliability of mine fire monitoring is improved.

Description

Technical Field

[0001] This invention belongs to the field of coal mine safety monitoring technology and relates to a method for suppressing false fire interference in coal mines based on time-series consistency analysis. Background Technology

[0002] Underground coal mine operations face severe fire risks, with gas combustion and spontaneous combustion of coal being the most common hazards. Flames release significant infrared spectral radiation and exhibit dynamic scintillation characteristics during combustion, making infrared spectral imaging a core tool for coal mine safety monitoring. By capturing thermal radiation signals, infrared spectral imaging can identify potential flame areas in real time, providing crucial information for fire early warning.

[0003] However, interference from non-flame light sources in coal mine environments is becoming increasingly prominent. For example, welding sparks from electric welding operations, sparks caused by mechanical friction, and electric arcs generated by electrical equipment discharges all exhibit bright spot characteristics similar to real flames in infrared spectral images. These interferences significantly increase the false alarm rate of traditional flame detection systems, not only reducing monitoring accuracy but also potentially leading to resource waste and operational risks under complex backgrounds and dynamic lighting conditions.

[0004] In real-world mine environments, the background often contains significant clutter and dynamic shadows, while lighting conditions change frequently due to equipment operation. Traditional detection methods rely on the spatial features of single-frame images for analysis, failing to distinguish between transient interference signals and persistent, real flames. This limitation exposes the inadequacy of existing technologies in terms of robustness and adaptability, making it difficult to meet the demands of high-precision flame identification. Therefore, developing an innovative method that can effectively suppress false-flame interference while maintaining flame detection sensitivity has become an urgent challenge for improving the reliability of coal mine safety monitoring systems. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a method for suppressing false fire interference in coal mine flames based on temporal consistency analysis, so as to solve the problem of false alarms caused by false fire sources such as electric arcs and welding sparks in existing flame detection methods, and improve the accuracy and robustness of flame identification.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for suppressing false fire interference in coal mines based on temporal consistency analysis includes the following steps:

[0008] S1: Obtain the infrared spectral flame saliency probability map of continuous image frames in underground coal mines;

[0009] S2: Based on the probability map, extract the flame candidate region. The extraction process includes using the Otsu adaptive threshold segmentation method to determine the segmentation threshold and combining it with morphological optimization and hole filling processing.

[0010] S3: Perform temporal consistency analysis on the flame candidate regions of consecutive image frames, and calculate the spatial overlap, morphological stability and infrared thermal intensity variation trend of adjacent frames.

[0011] S4: When the spatial overlap and thermal intensity change meet the preset threshold conditions, the candidate region is determined to be a real flame region; otherwise, it is determined to be an interference signal and is eliminated.

[0012] Furthermore, the formula for calculating the spatial overlap is:

[0013]

[0014] Among them, M t Indicates the flame mask region of the current frame; M t-1 Indicates the flame mask area of ​​the previous frame; O t Indicates the degree of spatial overlap matching.

[0015] Furthermore, the trend of infrared thermal intensity change is obtained by calculating the time series difference of the average gray value of the flame area. When the difference value is less than a preset threshold, it is determined that the thermal intensity change is stable.

[0016] Furthermore, the optimal threshold for the Otsu adaptive threshold segmentation satisfies the following equation:

[0017]

[0018] in, It is represented as the inter-class variance, and T is the threshold that maximizes the inter-class variance.

[0019] Furthermore, the morphological optimization includes erosion and dilation operations to remove noise points and smooth out boundary discontinuities.

[0020] Furthermore, the void-filling process is used to fill the internal voids in the flame area to generate a flame mask area with a complete shape.

[0021] Furthermore, the method is integrated into the infrared spectroscopy flame detection system as a separate post-processing module.

[0022] Furthermore, the method is used in conjunction with an infrared spectroscopy flame detection neural network, which is based on multi-level feature fusion to improve the accuracy and robustness of flame identification in complex mining environments.

[0023] The beneficial effects of this invention are as follows:

[0024] (1) By using the time-series consistency analysis method, it is possible to effectively distinguish between real flames and instantaneous bright spots such as electric arcs and welding sparks, and significantly reduce the false alarm rate;

[0025] (2) The method is more robust to image background complexity and illumination interference;

[0026] (3) It can be used as an independent post-processing module and directly embedded into the existing infrared spectroscopy flame detection system without significantly modifying the original detection architecture.

[0027] (4) It can also be used in conjunction with a deep learning-based flame detection neural network to achieve dual protection for flame recognition in complex coal mine environments.

[0028] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0029] To make the objectives, technical solutions, and advantages of the present invention clearer, the preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings, wherein:

[0030] Figure 1 This is a schematic flowchart of the method of the present invention;

[0031] Figure 2 This is a schematic diagram of the timing consistency analysis module of the present invention;

[0032] Figure 3 This is a schematic diagram of the present invention applied to an infrared flame detection system in a coal mine. Detailed Implementation

[0033] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and various details in this specification can be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0034] The accompanying drawings are for illustrative purposes only and are schematic diagrams, not actual pictures. They should not be construed as limiting the invention. To better illustrate the embodiments of the invention, some parts in the drawings may be omitted, enlarged, or reduced, and do not represent the actual product dimensions. It is understandable to those skilled in the art that some well-known structures and their descriptions may be omitted in the drawings.

[0035] In the accompanying drawings of the embodiments of the present invention, the same or similar reference numerals correspond to the same or similar components. In the description of the present invention, it should be understood that if terms such as "upper," "lower," "left," "right," "front," and "rear" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, they are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the terms used to describe positional relationships in the drawings are only for illustrative purposes and should not be construed as limiting the present invention. For those skilled in the art, the specific meaning of the above terms can be understood according to the specific circumstances.

[0036] like Figure 1 As shown, the coal mine flame false fire interference suppression method based on time-series consistency analysis provided by this invention mainly includes four steps:

[0037] S1: Acquisition of saliency probability map: Continuous image frames are obtained from the coal mine infrared image acquisition module, and an infrared spectral flame detection network is used to generate a flame saliency probability map.

[0038] S2: Flame candidate region extraction. The Otsu adaptive threshold segmentation algorithm is used on the saliency probability map to obtain flame candidate regions. Morphological operations and hole filling are combined to improve the integrity of the regions.

[0039] S3: Temporal consistency analysis, which calculates the spatial overlap, morphological stability and infrared thermal intensity variation trend of candidate regions in adjacent frames;

[0040] S4: Judgment and False Fire Suppression. When the spatial overlap of consecutive frames is higher than the threshold and the heat intensity changes steadily, it is judged as a real flame; otherwise, it is regarded as a false fire signal and is rejected.

[0041] like Figure 2 As shown, the timing consistency analysis module of the present invention includes:

[0042] Spatial overlap calculation unit, used for formula-based calculation

[0043] The morphological stability detection unit is used to detect whether the morphology of candidate regions in consecutive frames remains consistent.

[0044] The infrared thermal intensity change detection unit is used to calculate the difference in average gray value of candidate regions in consecutive frames. When the difference is lower than the threshold, the thermal intensity is determined to be stable.

[0045] The determination unit is used to comprehensively consider the above indicators to ultimately confirm the flame area or eliminate false flame areas.

[0046] like Figure 3 As shown, the method of the present invention can be integrated into a coal mine infrared spectroscopy flame detection system, including:

[0047] Infrared image acquisition module, used to acquire continuous infrared spectral images of underground coal mines;

[0048] The flame saliency detection module performs preliminary saliency analysis on the infrared image and generates a flame saliency probability map.

[0049] The temporal consistency analysis module receives candidate flame regions and uses spatial overlap, morphological stability, and thermal intensity changes to determine the actual flame.

[0050] The false fire suppression result output module outputs the suppressed flame detection result;

[0051] The host computer monitoring system receives and displays flame detection information and can be linked with the mine's automatic fire extinguishing device.

[0052] Example 1

[0053] based on Figure 1 The process described above involves acquiring a continuous image sequence (e.g., 100 frames) using an infrared camera at 25 frames per second in an underground flame detection experiment. First, a pre-trained infrared spectral flame detection network (such as a ResNet architecture) is used to generate a flame saliency probability map, with output probability values ​​ranging from 0 to 1. Then, the optimal segmentation threshold (typically T = 0.4) is determined using the Otsu adaptive thresholding algorithm, combined with morphological optimization (including 3×3 kernel erosion to remove noise points and 5×5 kernel dilation to smooth boundaries) and hole filling (using a region growing algorithm to fill internal holes) to obtain a complete flame candidate region. Subsequently, temporal consistency analysis was performed on the candidate regions in consecutive frames: when the spatial overlap (calculation formula is given in claim 2) is higher than the preset threshold of 0.6 (this threshold is set based on historical false alarm data analysis and can effectively distinguish between stable flames and transient interference), and the infrared thermal intensity difference (calculated as the difference between the average gray values ​​of adjacent frames) is less than the set threshold ΔT = 10 (to avoid misjudgment caused by illumination fluctuations), the true flame region is finally output; experimental data show that the false alarm rate is reduced by about 35% under this setting.

[0054] Example 2

[0055] based on Figure 2The temporal consistency analysis module, as shown in the diagram, can be implemented in an embedded system (such as an ARM Cortex-A53 platform) or a software environment (such as Python + OpenCV). The spatial overlap calculation unit is implemented through region mask comparison (specifically, binary mask intersection-union calculation using OpenCV's bitwise_and and bitwise_or functions); the morphological stability detection unit is implemented through morphological contour matching (using Canny edge detection to extract contours and calculating contour similarity using the Hu moment algorithm, with a threshold of 0.8); and the infrared thermal intensity change detection unit is implemented through time series analysis (using a sliding window method to calculate the difference in the average grayscale value sequence of 5 consecutive frames, with a difference threshold of 5). Finally, the judgment unit (based on logic and rules: spatial overlap > 0.6, morphological similarity > 0.8, and thermal intensity difference < 5) integrates the three results to output a flame authenticity judgment. Figure 2 The architecture shown was deployed on an NVIDIA Jetson Nano hardware platform during testing, with a processing latency of less than 50ms.

[0056] Example 3

[0057] like Figure 3 As shown, the method of this invention can be seamlessly integrated into existing coal mine infrared spectroscopy flame detection systems as a lightweight post-processing module (code size < 500 lines). The integration steps include: receiving the probability map (JSON format, including frame sequence coordinates and probability values) output by the flame saliency detection module via a RESTful API interface; running a sequence consistency analysis algorithm in an edge computing device (such as an industrial PLC); after false fire interference is eliminated (e.g., the success rate of removing welding spark interference signals is >90% in simulation tests), the final result is uploaded to a host computer monitoring system (such as a SCADA platform) via the Modbus protocol, achieving real-time flame monitoring and false alarm suppression. Actual deployment data shows that the overall false alarm rate of the system has decreased from 15% to 5%, and the response time remains within 200ms.

[0058] In the candidate region, the thermal intensity index is calculated by taking the average gray value of all pixels within the region (formula: Where N is the number of pixels, G i The grayscale value is obtained. Trend analysis uses the time-series differencing method: calculating the absolute value of the difference between the average grayscale values ​​of adjacent frames. If |ΔG| is in 5 consecutive frames t If |ΔG < 5 (the threshold is set based on mine thermal radiation stability experiments) and the standard deviation is less than 2, then the trend of change is considered to be gradual, and the flame can be considered to persist; if |ΔG t If the value is greater than 20 and the duration is less than 3 frames (e.g., a typical characteristic of arc interference), it is determined to be a sudden change in thermal intensity, which is very likely to be an interference signal and is automatically eliminated.

[0059] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for suppressing false fire interference in coal mines based on temporal consistency analysis, characterized in that: Includes the following steps: S1: Obtain the infrared spectral flame saliency probability map of continuous image frames in underground coal mines; S2: Based on the probability map, extract the flame candidate region. The extraction process includes using the Otsu adaptive threshold segmentation method to determine the segmentation threshold and combining it with morphological optimization and hole filling processing. S3: Perform temporal consistency analysis on the flame candidate regions of consecutive image frames, and calculate the spatial overlap, morphological stability and infrared thermal intensity variation trend of adjacent frames. S4: When the spatial overlap and thermal intensity change meet the preset threshold conditions, the candidate region is determined to be a real flame region; otherwise, it is determined to be an interference signal and is eliminated.

2. The method for suppressing false fire interference in coal mines based on temporal consistency analysis according to claim 1, characterized in that: The formula for calculating the spatial overlap is: Among them, M t Indicates the flame mask region of the current frame; M t-1 Indicates the flame mask area of ​​the previous frame; O t Indicates the degree of spatial overlap matching.

3. The method for suppressing false fire interference in coal mines based on temporal consistency analysis according to claim 1, characterized in that: The trend of infrared thermal intensity change is obtained by calculating the time series difference of the average gray value of the flame area. When the difference value is less than a preset threshold, it is determined that the thermal intensity change is stable.

4. The method for suppressing false fire interference in coal mines based on temporal consistency analysis according to claim 1, characterized in that: The optimal threshold for Otsu adaptive threshold segmentation satisfies the following formula: in, It is represented as the inter-class variance, and T is the threshold that maximizes the inter-class variance.

5. The method for suppressing false fire interference in coal mines based on temporal consistency analysis according to claim 1, characterized in that: The morphological optimization includes erosion and dilation operations to remove noise points and smooth out boundary discontinuities.

6. The method for suppressing false fire interference in coal mines based on temporal consistency analysis according to claim 1, characterized in that: The cavity filling process is used to fill the internal cavities in the flame area to generate a flame mask area with a complete shape.

7. The method for suppressing false fire interference in coal mines based on temporal consistency analysis according to claim 1, characterized in that: The method is integrated into the infrared spectroscopy flame detection system as a standalone post-processing module.

8. The method for suppressing false fire interference in coal mines based on temporal consistency analysis according to claim 1, characterized in that: The method is used in conjunction with an infrared spectroscopy flame detection neural network, which is based on multi-level feature fusion to improve the accuracy and robustness of flame identification in complex mining environments.

Images