Gas Leak Location Detection Method and Device

Abstract

The embodiments of this specification provide a method and device for locating the gas leakage position. The gas leakage position locating method includes: collecting spectral images corresponding to multiple bands for the gas leakage area by using a drone, fusing the spectral images corresponding to the multiple bands to obtain a multi-spectral image; extracting the multi-spectral features of the multi-spectral image, and constructing a gas concentration heat map corresponding to the gas leakage area according to the multi-spectral features; performing concentration image feature recognition on the gas concentration heat map, and determining an abnormal image area in the gas concentration heat map according to the recognition result; constructing a flight path for the drone in the gas leakage area according to the abnormal image area, and driving the drone to locate the gas leakage position based on the flight path.

Description

Technical Field

[0001] The embodiments of this specification relate to the field of image processing technology, and particularly to a method and device for locating the position of gas leakage. Background Art

[0002] Chemicals widely used in industrial production processes, especially flammable and toxic gas chemicals, pose potential threats to the environment and human health. In industries such as petrochemical, fertilizer manufacturing, and metallurgy, the leakage or excessive emission of toxic gases will cause serious air pollution. Therefore, toxic gas monitoring technology has emerged. Existing toxic gas monitoring technologies mainly rely on fixed sensor network systems. Such sensors are usually installed at specific locations in factories and storage areas, and monitor the concentration of specific gases in the air through a predetermined sampling frequency. When the gas concentration exceeds the preset threshold, the system issues an alarm. Although this method can achieve the monitoring purpose, it has certain limitations. First, fixed sensors can only cover a limited space and cannot monitor the entire factory or storage area in real time, resulting in monitoring blind spots in some areas. Second, the deployment and maintenance costs of fixed sensor systems are relatively high. Especially in harsh environments, sensors are prone to damage or failure, thereby reducing the reliability and effectiveness of monitoring. Therefore, an effective solution is urgently needed to solve the above problems. Summary of the Invention

[0003] In view of this, the embodiments of this specification provide a method for locating the position of gas leakage. One or more embodiments of this specification also relate to a device for locating the position of gas leakage, a computing device, a computer-readable storage medium, and a computer program product to solve the technical defects existing in the prior art.

[0004] According to the first aspect of the embodiments of this specification, a method for locating the position of gas leakage is provided, including:

[0005] Collect spectral images corresponding to multiple bands for the gas leakage area through a drone, and fuse the spectral images corresponding to the multiple bands to obtain a multi-spectral image;

[0006] Extract the multi-spectral features of the multi-spectral image, and construct a gas concentration heat map corresponding to the gas leakage area according to the multi-spectral features;

[0007] Perform concentration image feature recognition on the gas concentration heat map, and determine an abnormal image area in the gas concentration heat map according to the recognition result;

[0008] Construct a flight path for the drone in the gas leakage area according to the abnormal image area, and drive the drone to locate the gas leakage position based on the flight path.

[0009] According to the second aspect of the embodiments of this specification, a gas leakage position locating device is provided, including:

[0010] An acquisition module, configured to collect spectral images corresponding to multiple bands for a gas leakage area through a drone, and fuse the spectral images corresponding to the multiple bands to obtain a multispectral image;

[0011] An extraction module, configured to extract multispectral features of the multispectral image, and construct a gas concentration heat map corresponding to the gas leakage area according to the multispectral features;

[0012] An identification module, configured to perform concentration image feature identification on the gas concentration heat map, and determine an abnormal image area in the gas concentration heat map according to the identification result;

[0013] A positioning module, configured to construct a flight path for the drone in the gas leakage area according to the abnormal image area, and drive the drone to locate the gas leakage position based on the flight path.

[0014] According to the third aspect of the embodiments of this specification, a computing device is provided, including:

[0015] A memory and a processor;

[0016] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the above gas leakage position locating method are implemented.

[0017] According to the fourth aspect of the embodiments of this specification, a computer-readable storage medium is provided, which stores computer-executable instructions. When the instructions are executed by a processor, the steps of the above gas leakage position locating method are implemented.

[0018] According to the fifth aspect of the embodiments of this specification, a computer program product is provided, including a computer program or instructions. When the computer program or instructions are executed by a processor, the steps of the above gas leakage position locating method are implemented.

[0019] The gas leakage location positioning method provided in this embodiment can quickly and accurately locate the gas leakage location and is not affected by spatial positions. It can collect spectral images corresponding to multiple bands for the gas leakage area through a drone, and then fuse the spectral images corresponding to multiple bands to obtain a multispectral image. The multispectral image can carry the change information of gas concentrations in different regions of the air. Therefore, the multispectral features of the multispectral image can be extracted, and a gas concentration heat map corresponding to the gas leakage area can be constructed based on the multispectral features. The gas concentration heat map can intuitively reflect the gas changes in different regions. In order to quickly locate the gas leakage location, the concentration image features of the gas concentration heat map can be recognized, and the abnormal image area can be determined in the gas concentration heat map according to the recognition result. Then, the flight path of the drone can be constructed in the gas leakage area according to the abnormal image area, and the drone can be driven to continue positioning based on the flight path, so as to locate the gas leakage location. This realizes the rapid detection of the gas leakage location in a way that does not require installing sensors at fixed positions, and at the same time can avoid the risks brought by manual detection, improving the safety and flexibility of gas leakage detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a flowchart of a gas leakage location positioning method provided by an embodiment of this specification;

[0021] Figure 2 is a schematic diagram of image acquisition in a gas leakage location positioning method provided by an embodiment of this specification;

[0022] Figure 3 is a processing procedure flowchart of a gas leakage location positioning method provided by an embodiment of this specification;

[0023] Figure 4 is a schematic structural diagram of a gas leakage location positioning device provided by an embodiment of this specification;

[0024] Figure 5 is a structural block diagram of a computing device provided by an embodiment of this specification. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] Many specific details are set forth in the following description in order to provide a thorough understanding of this specification. However, this specification can be implemented in many other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the connotation of this specification. Therefore, this specification is not limited by the specific embodiments disclosed below.

[0026] The terms used in one or more embodiments of this specification are for the purpose of describing specific embodiments only and are not intended to limit one or more embodiments of this specification. The singular forms "a", "the", and "said" used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms unless the context clearly dictates otherwise. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and encompasses any and all possible combinations of one or more of the associated listed items.

[0027] It should be understood that although the terms first, second, etc. may be used in one or more embodiments of this specification to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "when" or "while" or "in response to determining".

[0028] In addition, it should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in one or more embodiments of this specification are all information and data that have been authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant laws, regulations, and standards of relevant countries and regions, and corresponding operation entrances are provided for users to choose to authorize or refuse.

[0029] First, the noun terms involved in one or more embodiments of this specification are explained.

[0030] Linear Discriminant Analysis (LDA, full name Linear Discriminant Analysis) is a classic statistical learning and data analysis technique. The core idea of LDA is to project data into a new space through a linear transformation so that samples of the same category are as compact as possible, while samples of different categories are as separated as possible. Specifically, LDA aims to find a linear combination of features to maximize the ratio of between-class variance to within-class variance, that is, the Fisher criterion. To achieve this goal, LDA needs to calculate the within-class scatter matrix (reflecting the dispersion degree between samples of the same class) and the between-class scatter matrix (reflecting the differences between samples of different classes), and find a straight line or hyperplane that can best project.

[0031] In this specification, a method for locating the gas leakage position is provided. This specification also relates to a gas leakage position locating device, a computing device, a computer-readable storage medium, and a computer program product, which will be described in detail one by one in the following embodiments.

[0032] In practical applications, to address the limitations of fixed sensors, some portable and mobile detection devices have emerged, such as handheld gas detectors. Such devices can be carried by staff for detection at any time and place. However, this method also has certain drawbacks. Due to relying on human labor, handheld detectors are difficult to meet the requirements of high-frequency and wide-range real-time monitoring. Especially in high-risk environments, frequent exposure to toxic gases may endanger the health and safety of detection personnel. In addition, the detection range of handheld devices is small, and it is difficult to generate a gas concentration distribution map of the area in real time. To achieve more efficient wide-area monitoring, vehicle-mounted mobile detection systems are also used in traditional methods. Such systems are installed on special vehicles and use the movement of the vehicle to detect gases in the area. However, the flexibility of vehicle-mounted systems is poor, and they cannot cover areas with complex terrains, especially the interior of factories with many height restrictions or multi-layer structures. In addition, the moving speed of the vehicle is limited, and it is impossible to achieve fine-grained and dynamic real-time monitoring. Therefore, in large-area and high-frequency toxic gas monitoring scenarios, traditional fixed sensors, handheld detectors, and vehicle-mounted detection systems all have deficiencies. Especially in the case where the leakage source needs to be quickly and accurately located, the limitations of these systems are more obvious. Therefore, an effective solution is urgently needed to solve the above problems.

[0033] The gas leakage position locating method provided in this embodiment, in order to quickly and accurately locate the gas leakage position and be unaffected by spatial positions, can collect spectral images corresponding to multiple bands for the gas leakage area through an unmanned aerial vehicle (UAV). Then, fuse the spectral images corresponding to multiple bands to obtain a multi-spectral image. The multi-spectral image can carry the change information of gas concentrations in different areas of the air. Therefore, the multi-spectral features of the multi-spectral image can be extracted, and a gas concentration heat map corresponding to the gas leakage area can be constructed based on the multi-spectral features. The gas concentration heat map can visually reflect the gas changes in different areas. In order to quickly locate the gas leakage position, the concentration image features of the gas concentration heat map can be further identified, and the abnormal image area can be determined in the gas concentration heat map according to the recognition result. Then, the flight path of the UAV can be constructed in the gas leakage area according to the abnormal image area, and the UAV can be driven to continue positioning based on the flight path, thereby locating the gas leakage position. This realizes the rapid detection of the gas leakage position by means of installing sensors without a fixed position, and at the same time can avoid the risks brought by manual detection, improving the safety and flexibility of gas leakage detection.

[0034] See Figure 1 , Figure 1 which shows a flowchart of a gas leakage location positioning method provided according to an embodiment of this specification, specifically including the following steps.

[0035] Step S102: Use a drone to collect spectral images corresponding to multiple bands for the gas leakage area, and fuse the spectral images corresponding to the multiple bands to obtain a multi-spectral image.

[0036] The gas leakage location positioning method provided in this embodiment can be applied to the positioning of the source location in any gas leakage scenario, such as the leakage of toxic gases in high-risk environments such as industrial areas and storage areas, or the leakage of gas in residential areas, or the leakage of toxic gases in laboratories, etc. This embodiment takes the leakage of toxic gases in an industrial area as an example to illustrate the gas leakage location positioning method provided in this embodiment. Descriptions of other scenarios can refer to the same or corresponding descriptions in this embodiment, and this embodiment does not make any limitations here.

[0037] It should be noted that the gas leakage location positioning method provided in this embodiment can not only be used for the leakage of toxic gases, but also for the leakage location positioning of gases stored in any enclosed space, so as to avoid the impact of the leaked gas on the environment.

[0038] Specifically, the gas leakage area specifically refers to areas such as industrial areas, storage areas, and residential areas where there is gas leakage that pollutes the environment. Correspondingly, the multiple bands specifically refer to the bands for which the drone collects spectral images of the gas leakage area, and the imaging effects of the spectral images collected in different bands are different. The spectral image specifically refers to the image obtained by collecting images of the gas leakage area through the spectral sensor configured on the drone. In practical applications, for different gases, different sensors can be selected in combination with their characteristic spectral writing bands. Correspondingly, the multi-spectral image specifically refers to the spectral image corresponding to the gas leakage area obtained after fusing the spectral images corresponding to the multiple bands, which contains multiple band information and can improve the positioning accuracy of the gas leakage location.

[0039] The gas leakage position localization method provided in this embodiment can quickly and accurately locate the gas leakage position and is not affected by the spatial position. Multiple spectral images corresponding to multiple bands can be collected for the gas leakage area by a drone, and then the spectral images corresponding to multiple bands are fused to obtain a multi-spectral image. The multi-spectral image can carry the change information of the gas concentration in different regions of the air. Therefore, the multi-spectral features of the multi-spectral image can be extracted, and a gas concentration heat map corresponding to the gas leakage area can be constructed based on the multi-spectral features. The gas concentration heat map can intuitively reflect the gas changes in different regions. In order to quickly locate the gas leakage position, the concentration image features of the gas concentration heat map can be recognized, and the abnormal image area can be determined in the gas concentration heat map according to the recognition result. Then, the flight path of the drone can be constructed in the gas leakage area according to the abnormal image area, and the drone can be driven to continue positioning based on the flight path, so as to locate the gas leakage position. In this way, the rapid detection of the gas leakage position can be completed by means of installing sensors without a fixed position, and at the same time, the risk brought by manual detection can be avoided, and the safety and flexibility of gas leakage detection are improved.

[0040] Further, when collecting spectral images by a drone and fusing them to construct a multi-spectral image, in order to ensure that the fused multi-spectral image can be used for subsequent gas leakage position localization, an image registration strategy can be adopted for image fusion processing. In this embodiment, the specific implementation method is as follows:

[0041] The multi-spectral sensor of the drone collects images of the gas leakage area at a set angle to obtain spectral images corresponding to multiple bands respectively; the spectral images corresponding to multiple bands are fused according to the image registration strategy to obtain a multi-spectral image.

[0042] Specifically, the multi-spectral sensor can be selected according to the characteristic spectral absorption bands of toxic gases (such as CO, H2S, etc.). Correspondingly, the image registration strategy specifically refers to the strategy of matching and superimposing two or more images obtained at different times, by different sensors (imaging devices), or under different conditions (weather, illumination, camera position and angle, etc.). The purpose is to find the spatial mapping relationship between the pixels of one image and the pixels of another image, so as to ensure that the multi-spectral image has a true mapping relationship with the gas leakage area.

[0043] Based on this, in order to ensure that the gas leakage position can be detected according to the multi-spectral image in the future, the multi-spectral sensor of the drone can collect images of the gas leakage area at a set angle to obtain spectral images corresponding to multiple bands respectively; then the spectral images corresponding to multiple bands can be fused according to the image registration strategy to obtain a multi-spectral image for subsequent use.

[0044] In practical applications, drones can also be equipped with gas detection sensors to detect gases in the shortwave infrared (SWIR) and ultraviolet (UV) bands, determining gas concentrations through differences in absorption and scattering for subsequent thermal mapping. When installing a multispectral camera on a drone, it is important to ensure stable camera shooting and avoid vibration interference. Therefore, a gimbal-mounted system can be used to ensure clear image capture during inspections. Furthermore, ground calibration is required to ensure the consistency and accuracy of multispectral images captured at different heights and angles.

[0045] Furthermore, once configured, the drone can take off from a corner of an area to capture spectral images of the gas leak location. The drone can then scan its surroundings, using its multispectral sensor to capture images at fixed angles, capturing multispectral images of its surroundings. Multispectral sensors typically capture images in different wavelengths, generating multiple images simultaneously at different wavelengths (such as near-infrared and ultraviolet). These images can then be fused into a complete spectral image using image registration strategies, allowing the data from each spectral channel to comprehensively represent the concentration distribution of the toxic gas. This image can then be used to locate the gas leak.

[0046] For example, a toxic gas detection device in an industrial area issues an alarm, indicating that a sealed tank containing toxic gas is leaking. In order to quickly locate the sealed tank where the gas leak occurs, a drone can be used to locate the gas leak. In this case, a drone equipped with a spectral sensor can be used to collect spectral images of various areas in the industrial area where gas may leak, such as Figure 2 As shown in the schematic diagram, spectral images corresponding to different bands are acquired through acquisition. These images can then be fused using image registration strategies to generate a multispectral image, which can be used to visualize the distribution of toxic gas concentrations for subsequent positioning.

[0047] In summary, by collecting spectral images of different bands by drones and fusing them to construct multispectral images, the multispectral images can carry the concentration distribution information of toxic gases, which makes it easier to locate the gas leak through subsequent image analysis.

[0048] Step S104 : extracting multispectral features of the multispectral image, and constructing a gas concentration heat map corresponding to the gas leakage area according to the multispectral features.

[0049] Specifically, after obtaining the multi-spectral image that records the distribution information of toxic gases corresponding to different bands, in order to facilitate the subsequent positioning of the gas leakage location, the multi-spectral features of the multi-spectral image can be extracted first, and then a gas concentration heat map corresponding to the gas leakage area can be constructed according to the multi-spectral features, so as to reflect the gas concentration corresponding to different areas through the gas concentration heat map. Furthermore, the UAV can be controlled to move towards the area with high concentration according to the gas concentration, so as to quickly and accurately locate the gas leakage location.

[0050] Among them, the multi-spectral image feature specifically refers to the vector expression obtained after feature extraction of the multi-spectral image. Correspondingly, the gas concentration heat map specifically refers to the heat map that reflects the gas concentration in different areas, and in this map, the concentration level can be characterized by different colors. For example, a certain area displayed in red indicates high concentration, and displayed in blue indicates low concentration, etc., so as to clearly visualize the gas distribution.

[0051] Furthermore, when constructing the gas concentration heat map, considering that this map can show the gas concentration at different positions within the area, the trained regression model can be combined to complete the concentration distribution prediction and map it to the spectral image to obtain the heat map. In this embodiment, the specific implementation method is as follows:

[0052] Process the multi-spectral image using the linear discriminant algorithm to obtain multi-spectral image features; input the multi-spectral image features into the regression model for processing to obtain the distribution information of the gas concentration values corresponding to the gas leakage area; map the distribution information of the gas concentration values to the multi-spectral image, and generate the gas concentration heat map corresponding to the gas leakage area according to the mapping result.

[0053] Specifically, the linear discriminant algorithm specifically refers to the LDA algorithm. Correspondingly, the regression model specifically refers to the trained model that can calculate the gas concentration value corresponding to each area according to the multi-spectral image features. That is to say, the regression model has learned the mapping relationship between the gas concentration and the spectral image features. After inputting the multi-spectral image features, the model can calculate the gas concentration value of each area according to the learned knowledge and output it for constructing the gas concentration heat map. Correspondingly, the distribution information of the gas concentration values specifically refers to the distribution information corresponding to all the gas concentration values obtained by calculating the gas concentration value for each pixel point, which not only includes the gas concentration value but also the spatial distribution information.

[0054] Based on this, after obtaining the multispectral image, the linear discriminant algorithm can be used to process the multispectral image to obtain multispectral image features. Thereafter, in order to be able to map from the image to the gas concentration value and then construct a heat map, the multispectral image features can be input into a regression model for processing to obtain the gas concentration value distribution information corresponding to the gas leakage area. At this time, the gas concentration value distribution information can be mapped to the multispectral image, and thus a gas concentration heat map corresponding to the gas leakage area can be generated according to the mapping result.

[0055] In practical applications, by using the linear discriminant analysis (LDA) dimensionality reduction technique, significant features of the specified toxic gas in the multispectral image can be extracted. Through LDA, the best linear combination for distinguishing gas regions with different concentrations can be found. Therefore, after reducing the multispectral data to a small number of key features, the calculation efficiency can be significantly improved. On this basis, concentration estimation and heat map generation can be carried out. Specifically, the regression algorithm can be first applied to the extracted multispectral feature data to estimate the gas concentration value, and then it can be mapped to the spatial position of the image to generate a gas concentration heat map.

[0056] In specific implementation, considering that the gas concentration heat map can intuitively reflect the gas concentration in different regions and can be directly viewed by the corresponding personnel, pseudo-color coding can be adopted. For example, the high-concentration region is displayed in red and the low-concentration region is displayed in blue to clearly visualize the gas distribution. Thus, it is convenient for operation and maintenance personnel and fire fighters to directly observe the leakage point when discovering a suspicious leakage source, facilitating the subsequent implementation of security enhancement work.

[0057] Continuing with the above example, after obtaining the multispectral image, the LDA technique can be used to process the multispectral image, and thus the multispectral image features V corresponding to the multispectral image can be obtained. Thereafter, by combining the regression algorithm to process the multispectral image features V, the gas concentration value corresponding to each pixel point can be estimated. Then, mapping it to the multispectral image, a gas concentration heat map corresponding to the gas leakage area can be obtained for subsequent analysis of the map to quickly locate the gas leakage position.

[0058] In summary, by adopting the regression algorithm combined with the LDA technique to complete the construction of the heat map, it can be ensured that the heat map is more in line with the actual situation, thereby effectively improving the positioning accuracy and efficiency of the gas leakage position.

[0059] Step S106, perform concentration image feature recognition on the gas concentration heat map, and determine an abnormal image area in the gas concentration heat map according to the recognition result.

[0060] Specifically, after obtaining the gas concentration heat map corresponding to the gas leakage area, considering that the gas concentration heat map can represent the gas concentration values corresponding to each area, in order to quickly locate the gas leakage position, concentration image feature recognition can be performed on the gas concentration heat map. By performing feature recognition at the image level, it is possible to determine the abnormal image area in the gas concentration heat map according to the recognition result, so as to drive the drone to move in combination with the abnormal image area subsequently, and then quickly locate the gas leakage position.

[0061] Among them, concentration image feature recognition specifically refers to the recognition processing operation of recognizing the features representing gas concentration in the gas concentration heat map and determining the abnormal image area with the highest gas leakage risk. Correspondingly, the abnormal image area specifically refers to the image area corresponding to the pixel point with the highest gas concentration value in the gas concentration heat map, which can represent that the real geographical area corresponding to this image area has the highest gas leakage risk.

[0062] Furthermore, when determining the abnormal image area through image recognition, the gradient calculation method can be used to complete image detection and determine the abnormal image area. In this embodiment, the specific implementation method is as follows:

[0063] Perform gradient calculation on the gas concentration heat map to obtain the gradient change information between the gas concentration value and the heat map pixel points in the gas concentration heat map; detect the gas concentration heat map according to the preset sliding window according to the gradient change information, and determine the abnormal image area in the gas concentration heat map according to the detection result.

[0064] Specifically, the gradient change information between the gas concentration value and the heat map pixel points specifically refers to the trend information of the gas concentration value changing with position, which is used for abnormal image area positioning. Correspondingly, the sliding window specifically refers to a window with a set size for detecting each pixel point in the gas concentration heat map.

[0065] Based on this, after obtaining the gas concentration heat map, considering that the gas concentration heat map can reflect the gas concentration values corresponding to different areas, therefore, at the image level, gradient calculation can be performed on the gas concentration heat map first to obtain the gradient change information between the gas concentration value and the heat map pixel points in the gas concentration heat map; at this time, the gas concentration heat map can be detected according to the preset sliding window according to the gradient change information, and the abnormal image area can be determined in the gas concentration heat map according to the detection result, so as to plan the flight path of the drone according to this area subsequently, and then quickly and accurately locate the gas leakage area.

[0066] Furthermore, in order to ensure the determination accuracy of the abnormal image area, it can be realized by determining and comparing the concentration values of each pixel point in the figure. In this embodiment, the specific implementation method is as follows:

[0067] Determine multiple candidate thermal image pixels that meet the gas leakage detection conditions in the gas concentration thermal map according to the detection results; determine the candidate gas concentration values corresponding to each candidate thermal image pixel, and compare the candidate gas concentration values with a preset concentration threshold; select the candidate thermal image pixels corresponding to the candidate gas concentration values greater than the concentration threshold according to the comparison results to construct a pixel set; perform weighted average calculation on the thermal image pixels included in the pixel set, and determine the abnormal image area in the gas concentration thermal map according to the calculation results.

[0068] Specifically, the gas leakage detection condition specifically refers to the condition for comparing the gas concentration values corresponding to each pixel in the gas concentration thermal map, and candidate thermal image pixels can be screened out according to the comparison results. Correspondingly, the candidate gas concentration value specifically refers to the pixel value corresponding to each candidate thermal image pixel.

[0069] Based on this, when locating the abnormal image area, considering that the gas concentration thermal map consists of an image and the gas concentration value corresponding to each pixel, after obtaining the gradient change information, pixel detection can be performed based on the sliding window according to the gradient change information. At this time, multiple candidate thermal image pixels that meet the gas leakage detection conditions can be screened out in the gas concentration thermal map according to the detection results; then, the candidate gas concentration values corresponding to each candidate thermal image pixel can be determined, and the candidate gas concentration values are compared with a preset concentration threshold; a pixel set is constructed by selecting the candidate thermal image pixels corresponding to the candidate gas concentration values greater than the concentration threshold; and weighted average calculation is performed on the thermal image pixels included in the pixel set, so as to determine the abnormal image area in the gas concentration thermal map according to the calculation results, which is convenient for subsequent path planning of the drone using the abnormal image area.

[0070] In practical applications, assume that after multi-spectral image acquisition of the gas leakage area by a drone, a two-dimensional thermal map I(x, y) including gas concentration is generated according to the acquired image, where (x, y) represents the position of the image pixel, and I(x, y) represents the gas concentration value at this position. On this basis, the gradient of the thermal map can be calculated to identify the trend of gas concentration change with position. The concentration gradient can be expressed by the following formula (1):

[0071] (1)

[0072] where the modulus of the gradient represents the rate of concentration change, and a larger gradient value indicates a significant concentration change.

[0073] After obtaining the concentration gradient, in order to identify the high-concentration regions in the heat map, a window with a size of w can be set, such as w = 5, and the pixel point with the local maximum gas concentration value in the heat map can be found within the window. For each pixel point (x, y) in the heat map, if its concentration value is I(x, y), the following condition of formula (2) needs to be satisfied:

[0074] (2)

[0075] Then at this time, (x, y) is the pixel point with the local maximum gas concentration value, that is, the area corresponding to this point may be the high-concentration peak area, which can be understood as the abnormal image area. Among them, i and j represent the values of half of the length and half of the width of the window.

[0076] Further, a preset concentration threshold T can be used to filter out the local peaks with low concentrations, so as to ensure that the detected pixel points are real high-concentration regions. If a pixel point with a local maximum gas concentration value satisfies I(x, y) > T, the concentration value corresponding to this pixel point can be marked as a high-concentration score.

[0077] Finally, considering that the pixel points corresponding to stable peaks can better reflect the true distribution of gas concentration, the weighted average position can be calculated based on a group of pixel points corresponding to high-concentration values, so as to locate the abnormal image area. Suppose the pixel points corresponding to the detected high-concentration values are , and its concentration value is Then the center point of the abnormal image area can be expressed by the following formula (3):

[0078] (3)

[0079] Based on the above processing, the central position corresponding to a smooth abnormal image area can be obtained, which can be used as an estimate of the high-concentration area at the current UAV position. By the above processing, the abnormal image area corresponding to the high-concentration peak can be determined in the gas concentration heat map, and the potential position of the gas leakage can be located. Subsequently, the UAV can send the real-time generated gas concentration heat map to the ground control center through wireless transmission through the central control system and save the data for subsequent analysis.

[0080] In specific implementation, considering that the drone collects images for the gas leakage area at set angles, in order to facilitate the positioning of the gas leakage location, the images collected from multiple angles can be stitched together to generate a panoramic concentration distribution map. When stitching, the phase correlation method image registration technology is used to align adjacent images spatially. At the same time, the concentration data in each multispectral image is superimposed on the panoramic image according to the angle and distance to generate a unified concentration heat map. Through this map, the surrounding concentration distribution can be displayed to help the drone locate the direction of high concentration. In the stitched panoramic concentration heat map, the area with the highest gas concentration (peak value) is identified. The position of this area corresponds to an angle, that is, the deflection angle starting from the front of the drone. The distance of the high concentration area is estimated through the depth information of the multispectral image and the height of the drone. This can be judged by calculating the gradient change rate of the concentration distribution map, that is, the high concentration area farther from the drone will appear more blurred in the image, while the concentration change in the closer area is more obvious.

[0081] Continuing with the above example, after obtaining the gas concentration heat map corresponding to the gas leakage area, the trend of the gas concentration changing with position can be calculated first through the above formula (1). Then, the gas concentration heat map can be detected according to this trend and in combination with a sliding window with a window size of 5, and the pixel points corresponding to the local maximum gas concentration value can be determined according to the detection result in combination with formula (2). Further, the gas concentration value corresponding to each pixel point can be compared with a preset concentration threshold T to screen out the pixel points with a high enough gas concentration value according to the comparison result and form a high-concentration point set. Thereafter, the pixel points included in the high-concentration point set can be weighted and averaged according to formula (3), and the central position coordinates (x 中 , y 中 ) can be determined according to the calculation result. On this basis, the high-concentration peak area can be determined in the gas concentration heat map according to the central position coordinates, indicating that there is a potential gas leakage risk in this area, so as to locate the gas leakage position in this area subsequently.

[0082] In summary, by using the method of image analysis and detection to determine the abnormal image area, the positioning accuracy of the abnormal image area can be guaranteed, and then it is convenient to drive the drone to move according to this area subsequently, so that the gas leakage position can be located quickly and accurately.

[0083] Step S108, construct a flight path for the drone in the gas leakage area according to the abnormal image area, and drive the drone to locate the gas leakage position based on the flight path.

[0084] Specifically, after determining the abnormal image area from the gas concentration heat map, considering that the abnormal image area is an area in the image and there is a risk of gas leakage in this area, a flight path can be constructed for the drone in the gas leakage area according to the abnormal image area, so that the drone can move towards the gas leakage position along this flight path and accurately locate the gas leakage position in the gas leakage area, so as to quickly carry out leakage treatment and avoid affecting the air.

[0085] Among them, the flight path specifically refers to a path planned for the flight direction, speed, altitude, angle, route, etc. of the drone, so that the drone can approach the gas leakage position along this path. Correspondingly, the gas leakage position specifically refers to the position where toxic gas leakage occurs in the gas leakage area.

[0086] Furthermore, after determining the abnormal image area, in order to enable the drone to quickly locate the gas leakage position, the drone needs to further monitor this area. Considering that there may be multiple abnormal image areas, in order to quickly locate the leakage position, a path planning method that matches the flight direction can be adopted. In this embodiment, the specific implementation method is as follows:

[0087] When there are multiple abnormal image areas in the gas concentration heat map, each abnormal image area is respectively matched with the flight direction of the drone; according to the matching result, the target abnormal image area is determined, and a flight path is constructed for the drone in the gas leakage area according to the target abnormal image area.

[0088] Specifically, the flight direction specifically refers to the direction in which the drone is currently moving forward. Correspondingly, the target abnormal image area specifically refers to the abnormal image area that matches the flight direction of the drone among multiple abnormal image areas, which can avoid multiple flight adjustments of the drone, quickly fly to the area corresponding to the target abnormal image area, and thus improve the positioning speed of the gas leakage position.

[0089] Based on this, when there are multiple abnormal image areas in the gas concentration heat map, if the drone is controlled to fly blindly, it may take a long time to locate the gas leakage position, resulting in more serious impacts. Therefore, each abnormal image area can be respectively matched with the flight direction of the drone to determine the target abnormal image area according to the matching result. After that, a flight path can be constructed for the drone in the gas leakage area according to the target abnormal image area, so as to prompt the drone to quickly locate the gas leakage position along the flight path.

[0090] Even further, in order to enable the drone to quickly and accurately complete subsequent monitoring and processing, the flight path planning can be carried out by calculating the image distance. In this embodiment, the specific implementation method is as follows:

[0091] Calculate the image distance between the abnormal image area and the UAV according to the gradient change information, and compare the image distance with a preset distance threshold;

[0092] When the image distance is less than the distance threshold, determine the original flight parameters corresponding to the UAV, and construct a flight path for the UAV in the gas leakage area based on the original flight parameters;

[0093] When the image distance is greater than or equal to the distance threshold, determine the flight parameters to be adjusted corresponding to the UAV; adjust the flight parameters to be adjusted according to the gradient change information and the image distance to obtain the target flight parameters; construct a flight path for the UAV in the gas leakage area based on the target flight parameters.

[0094] Specifically, the image distance specifically refers to the distance between the abnormal image area and the current position of the UAV in the image, and this distance calculation can be obtained according to the central position of the abnormal image area and the mapped position of the UAV in the image. Correspondingly, the original flight parameters specifically refer to the current unadjusted flight parameters of the UAV. Correspondingly, the flight parameters to be adjusted specifically refer to the flight parameters that can be adjusted by the UAV at the current moment. The target flight parameters are the flight parameters obtained after adjusting the flight parameters to be adjusted.

[0095] Based on this, when constructing a flight path, considering that different image distances require different path planning for the UAV, the image distance between the abnormal image area and the UAV can be calculated according to the gradient change information, and the image distance can be compared with a preset distance threshold.

[0096] When the image distance is less than the distance threshold, it indicates that the distance between the UAV and the position mapped by the abnormal image area in the actual space is relatively close at this time. Therefore, the original flight parameters corresponding to the UAV can be directly determined, and a flight path for the UAV can be constructed in the gas leakage area based on the original flight parameters, so as to prompt the UAV to quickly approach the position where gas may leak.

[0097] When the image distance is greater than or equal to the distance threshold, it indicates that the distance between the UAV and the position mapped by the abnormal image area in the actual space is relatively far at this time. In order to enable the UAV to quickly approach the gas leakage position, the flight parameters to be adjusted corresponding to the UAV can be determined; and the flight parameters to be adjusted can be adjusted according to the gradient change information and the image distance to obtain the target flight parameters; thereafter, a flight path for the UAV can be constructed in the gas leakage area based on the target flight parameters, so as to prompt the UAV to quickly approach the position where gas may leak.

[0098] In practical applications, when performing path planning for an unmanned aerial vehicle (UAV), considering that the abnormal image area has been determined as above, it is only necessary to locate the spatial area corresponding to this image area in the gas leakage area, so that the UAV can approach the gas leakage position. Therefore, it is assumed that the center point of this spatial area is located at . is the angle relative to the current orientation of the UAV. At this time, the UAV should move towards this angle to gradually approach the high-concentration area corresponding to the abnormal image area.

[0099] During this process, if there are multiple abnormal image areas in the gas concentration heat map, the high-concentration area corresponding to the abnormal image area closest to the UAV's flight direction can be selected as the area that the UAV needs to fly towards. Further, after determining the area to fly towards, the distance between this area and the UAV can be estimated using the concentration gradient change information of the gas concentration heat map. Specifically, during implementation, the gradient intensity can be calculated using the following formula (4):

[0100] (4)

[0101] Among them, areas with higher gradients usually indicate larger concentration changes and may be closer to the gas leakage position.

[0102] Based on this, the abnormal image area with the highest concentration can be identified in the gas concentration heat map through the concentration gradient intensity or the depth information of the spectral image. The pixel points corresponding to this area can be . After that, the distance between the position of this pixel point and the current position of the UAV can be calculated . If the concentration gradient changes greatly, it further indicates a shorter distance. Conversely, the smaller the change rate, the farther the distance.

[0103] When the distance to the high-concentration area is far, a larger step size (such as the standard flight speed of the UAV) can be selected to increase the flight speed. At the same time, as the UAV approaches the high-concentration area, the step size can be gradually reduced to enable the UAV to more accurately lock onto the center of the high-concentration area. Generally, the step size can be dynamically adjusted according to the concentration gradient and the distance, and their relationship is as follows formula (5):

[0104] (5)

[0105] Among them, k is an adjustment parameter that can take values between 0 and 1. Based on the above processing, the optimal flight path of the UAV at the current moment can be planned, enabling it to quickly approach the gas leakage position.

[0106] In summary, in order to enable the UAV to quickly approach the gas leakage location, the flight path can be constructed by combining the gas concentration heat map and the abnormal image area, so that the path planning of the UAV can be more reasonable and efficient.

[0107] On this basis, in each sampling period, it is necessary to locate the sampling position of the UAV and move it to a new position for spectral image acquisition, and then perform a new round of positioning processing. In this embodiment, the specific implementation method is as follows:

[0108] Drive the UAV to move in the gas leakage area based on the flight path; in the case where the position of the UAV after movement does not meet the positioning condition, calculate a new sampling position according to the heading angle of the UAV and the position of the UAV after movement; drive the UAV to move to the new sampling position, and execute the step of collecting spectral images corresponding to multiple bands for the gas leakage area through the UAV until the position of the UAV after movement meets the positioning condition, and determine the gas leakage position.

[0109] Based on this, after the path planning for the UAV is completed, the UAV can be driven to move in the gas leakage area based on the flight path; in the case where the position of the UAV after movement does not meet the positioning condition, it means that the position of the UAV at this time is not close enough to the gas leakage position, so it is necessary to control the UAV to move further, and then perform a new round of image acquisition and analysis. Therefore, a new sampling position can be calculated according to the heading angle of the UAV and the position of the UAV after movement; then the UAV can be driven to move to the new sampling position, and return to execute step S102; until the position of the UAV after movement meets the positioning condition, it can be explained that the current position of the UAV can locate the gas leakage position, so that the gas leakage position can be submitted to the operation and maintenance personnel to quickly take response measures, such as evacuating personnel, cutting off the leakage source, ventilation treatment, etc.

[0110] In practical applications, after the flight path for the UAV is constructed, the flight direction of the UAV can be adjusted according to the flight path, so that the UAV should be oriented at an angle towards the high-concentration area Move. The heading angle of the UAV is adjusted to so that it moves in the direction of the concentration peak in the next step. Therefore, the new position of the next sampling point of the UAV is calculated as the following formula (6):

[0111]

[0112] (6)

[0113] Based on the above-mentioned UAV flight control, the UAV can continuously narrow the detection range and gradually approach the high-concentration area. The flight direction and step size of each flight are dynamically adjusted based on the latest concentration heat map to ensure that the UAV always moves along the most effective path and finally reaches the location of the gas leakage source. In this way, real-time path optimization is achieved at each step, ensuring that the UAV effectively approaches the gas leakage source with the least number of detections, thereby achieving fast inspection efficiency and high-precision detection results.

[0114] Continuing with the above example, after determining the high-concentration peak area corresponding to the coordinates (x 中 , y 中 ), the distance d between the current position of the UAV and (x 中 , y 中 ) can be calculated. When the distance d is relatively far, the flight speed of the UAV needs to be increased to approach this area and enter the next image acquisition and processing. When the distance d is relatively close, the flight speed of the UAV can be gradually reduced so that the UAV can approach the gas leakage position more accurately. Through such iteration, it is finally determined that the sealed tank numbered 85 in the industrial area has a gas leak. At this time, the staff near the sealed tank 85 can be evacuated, and at the same time, the gas transmission pipeline of the sealed tank 85 can be cut off to avoid personnel and asset losses.

[0115] The gas leakage position positioning method provided in this embodiment, in order to quickly and accurately locate the gas leakage position and be unaffected by spatial positions, can collect spectral images corresponding to multiple bands for the gas leakage area through a UAV, and then fuse the spectral images corresponding to multiple bands to obtain a multi-spectral image; the multi-spectral image can carry the change information of gas concentrations in different regions of the air. Therefore, the multi-spectral features of the multi-spectral image can be extracted, and a gas concentration heat map corresponding to the gas leakage area can be constructed according to the multi-spectral features; the gas concentration heat map can intuitively reflect the gas changes in different regions. In order to quickly locate the gas leakage position, the concentration image features of the gas concentration heat map can be further identified, and the abnormal image area can be determined in the gas concentration heat map according to the identification result; then, the flight path for the UAV can be constructed in the gas leakage area according to the abnormal image area, and the UAV can be driven to continue positioning based on the flight path, thereby locating the gas leakage position. In this way, the fast detection of the gas leakage position is completed in a way that does not rely on installing sensors at fixed positions, and at the same time, the risks brought by manual detection can be avoided, improving the safety and flexibility of gas leakage detection.

[0116] The following Figure 3 , taking the application of the gas leakage position positioning method provided in this specification in the gas leakage scenario of coal gas as an example, further illustrates the gas leakage position positioning method. Among them,Figure 3 The figure shows a flowchart of the processing procedure of a gas leakage position locating method provided by an embodiment of this specification, which specifically includes the following steps.

[0117] Step S302: Use the multispectral sensor of the drone to collect images of the gas leakage area at a set angle, and obtain spectral images corresponding to multiple bands.

[0118] Step S304: Fusion the spectral images corresponding to multiple bands according to the image registration strategy to obtain a multispectral image.

[0119] Step S306: Process the multispectral image using the linear discriminant algorithm to obtain the features of the multispectral image.

[0120] Step S308: Input the features of the multispectral image into the regression model for processing to obtain the distribution information of the gas concentration values corresponding to the gas leakage area.

[0121] Step S310: Map the distribution information of the gas concentration values to the multispectral image, and generate a gas concentration heat map corresponding to the gas leakage area according to the mapping result.

[0122] Step S312: Calculate the gradient of the gas concentration heat map to obtain the gradient change information between the gas concentration values and the heat map pixel points in the gas concentration heat map.

[0123] Step S314: Detect the gas concentration heat map according to the preset sliding window according to the gradient change information, and determine the abnormal image area in the gas concentration heat map according to the detection result.

[0124] Step S316: Calculate the image distance between the abnormal image area and the drone according to the gradient change information, and compare the image distance with the preset distance threshold.

[0125] Step S318: When the image distance is less than the distance threshold, determine the original flight parameters corresponding to the drone, and construct a flight path for the drone in the gas leakage area based on the original flight parameters.

[0126] Step S320: Drive the drone to locate the gas leakage position based on the flight path, and upload the gas leakage position to the server.

[0127] In summary, in order to quickly and accurately locate the gas leakage position and be unaffected by the spatial position, spectral images corresponding to multiple bands can be collected for the gas leakage area by a drone, and then the spectral images corresponding to multiple bands are fused to obtain a multispectral image; the multispectral image can carry the change information of the gas concentration in different regions of the air. Therefore, the multispectral features of the multispectral image can be extracted, and a gas concentration heat map corresponding to the gas leakage area can be constructed according to the multispectral features; the gas concentration heat map can intuitively reflect the gas changes in different regions, and in order to quickly locate the gas leakage position. The concentration image features of the gas concentration heat map can be recognized, and the abnormal image area can be determined in the gas concentration heat map according to the recognition result; then the flight path of the drone can be constructed in the gas leakage area according to the abnormal image area, and the drone can be driven to continue positioning based on the flight path, so as to locate the gas leakage position. In this way, the rapid detection of the gas leakage position can be completed by means of disassembling the sensor installed at a fixed position, and at the same time, the risk brought by manual detection can be avoided, and the safety and flexibility of the gas leakage detection are improved.

[0128] Corresponding to the above method embodiment, this specification also provides an embodiment of a gas leakage position positioning device. Figure 4 The structural schematic diagram of a gas leakage position positioning device provided by an embodiment of this specification is shown. As Figure 4 shown, the device includes:

[0129] An acquisition module 402, configured to collect spectral images corresponding to multiple bands for the gas leakage area by a drone, and fuse the spectral images corresponding to the multiple bands to obtain a multispectral image;

[0130] An extraction module 404, configured to extract the multispectral features of the multispectral image, and construct a gas concentration heat map corresponding to the gas leakage area according to the multispectral features;

[0131] An identification module 406, configured to perform concentration image feature identification on the gas concentration heat map, and determine an abnormal image area in the gas concentration heat map according to the identification result;

[0132] A positioning module 408, configured to construct a flight path for the drone in the gas leakage area according to the abnormal image area, and drive the drone to locate the gas leakage position based on the flight path.

[0133] In an optional embodiment, the acquisition module 402 is further configured to:

[0134] The multi-spectral sensor of the drone collects images at a set angle for the gas leakage area to obtain spectral images corresponding to multiple bands; the spectral images corresponding to the multiple bands are fused according to an image registration strategy to obtain a multi-spectral image.

[0135] In an optional embodiment, the extraction module 404 is further configured to:

[0136] Process the multi-spectral image using a linear discriminant algorithm to obtain multi-spectral image features; input the multi-spectral image features into a regression model for processing to obtain the gas concentration value distribution information corresponding to the gas leakage area; map the gas concentration value distribution information to the multi-spectral image, and generate a gas concentration heat map corresponding to the gas leakage area according to the mapping result.

[0137] In an optional embodiment, the recognition module 406 is further configured to:

[0138] Perform gradient calculation on the gas concentration heat map to obtain the gradient change information between the gas concentration value and the heat map pixel points in the gas concentration heat map; detect the gas concentration heat map according to the preset sliding window according to the gradient change information, and determine the abnormal image area in the gas concentration heat map according to the detection result.

[0139] In an optional embodiment, the recognition module 406 is further configured to:

[0140] Determine multiple candidate heat map pixel points that meet the gas leakage detection conditions in the gas concentration heat map according to the detection result; determine the candidate gas concentration value corresponding to each candidate heat map pixel point, and compare the candidate gas concentration value with a preset concentration threshold; select the candidate heat map pixel points corresponding to the candidate gas concentration values greater than the concentration threshold according to the comparison result to construct a pixel point set; perform weighted average calculation on the heat map pixel points included in the pixel point set, and determine the abnormal image area in the gas concentration heat map according to the calculation result.

[0141] In an optional embodiment, the positioning module 408 is further configured to:

[0142] In the case where there are multiple abnormal image areas in the gas concentration heat map, match each abnormal image area with the flight direction of the drone respectively; determine the target abnormal image area according to the matching result, and construct a flight path for the drone in the gas leakage area according to the target abnormal image area.

[0143] In an optional embodiment, the positioning module 408 is further configured to:

[0144] Calculate the image distance between the abnormal image area and the drone according to the gradient change information, and compare the image distance with a preset distance threshold; when the image distance is less than the distance threshold, determine the original flight parameters corresponding to the drone, and construct a flight path for the drone in the gas leakage area based on the original flight parameters.

[0145] In an alternative embodiment, the positioning module 408 is further configured to:

[0146] When the image distance is greater than or equal to the distance threshold, determine the flight parameters to be adjusted corresponding to the drone; adjust the flight parameters to be adjusted according to the gradient change information and the image distance to obtain target flight parameters; construct a flight path for the drone in the gas leakage area based on the target flight parameters.

[0147] In an alternative embodiment, the positioning module 408 is further configured to:

[0148] Drive the drone to move in the gas leakage area based on the flight path; when the position after the drone moves does not meet the positioning condition, calculate a new sampling position according to the heading angle of the drone and the position after the drone moves; drive the drone to move to the new sampling position, and perform the step of collecting spectral images corresponding to multiple bands for the gas leakage area by the drone until the position after the drone moves meets the positioning condition, and determine the gas leakage position.

[0149] The gas leakage position locating device provided in this embodiment can quickly and accurately locate the gas leakage position and is not affected by the spatial position. It can collect spectral images corresponding to multiple bands for the gas leakage area through a drone, and then fuse the spectral images corresponding to multiple bands to obtain a multi-spectral image. The multi-spectral image can carry the change information of the gas concentration in different areas of the air. Therefore, the multi-spectral features of the multi-spectral image can be extracted, and a gas concentration heat map corresponding to the gas leakage area can be constructed based on the multi-spectral features. The gas concentration heat map can visually reflect the gas changes in different areas. In order to quickly locate the gas leakage position, the concentration image features of the gas concentration heat map can be recognized, and the abnormal image area can be determined in the gas concentration heat map according to the recognition result. Then, the flight path of the drone can be constructed in the gas leakage area according to the abnormal image area, and the drone can be driven to continue positioning based on the flight path, so as to locate the gas leakage position. In this way, the rapid detection of the gas leakage position can be completed by means of disassembling the sensor installed at a fixed position, and at the same time, the risk brought by manual detection can be avoided, and the safety and flexibility of the gas leakage detection are improved.

[0150] The above is a schematic solution of a gas leakage position locating device according to this embodiment. It should be noted that the technical solution of the gas leakage position locating device belongs to the same concept as the technical solution of the above gas leakage position locating method. For the details not described in the technical solution of the gas leakage position locating device, reference can be made to the description of the technical solution of the above gas leakage position locating method.

[0151] Figure 5 The structural block diagram of a computing device 500 provided according to an embodiment of this specification is shown. The components of the computing device 500 include, but are not limited to, a memory 510 and a processor 520. The processor 520 is connected to the memory 510 through a bus 530, and a database 550 is used to store data.

[0152] The computing device 500 also includes an access device 540, which enables the computing device 500 to communicate via one or more networks 560. Examples of such networks include the Public Switched Telephone Network (PSTN), Local Area Network (LAN), Wide Area Network (WAN), Personal Area Network (PAN), or a combination of communication networks such as the Internet. The access device 540 may include one or more of any type of wired or wireless network interfaces (e.g., a network interface card (NIC)), such as an IEEE 802.11 Wireless Local Area Network (WLAN) wireless interface, a Worldwide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, or a Near Field Communication (NFC) interface.

[0153] In one embodiment of the present specification, the above components of the computing device 500, as well as Figure 5 other components not shown, may also be connected to each other, for example, via a bus. It should be understood that Figure 5 the block diagram of the computing device shown is for illustrative purposes only and is not a limitation on the scope of the present specification. Those skilled in the art may add or replace other components as needed.

[0154] The computing device 500 can be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smartphone), a wearable computing device (e.g., a smartwatch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or a Personal Computer (PC). The computing device 500 can also be a mobile or stationary server.

[0155] Wherein, the processor 520 is configured to execute the following computer-executable instructions, which, when executed by the processor, implement the steps of the above gas leakage location positioning method.

[0156] The above is a schematic solution of a computing device according to this embodiment. It should be noted that the technical solution of this computing device and the technical solution of the above gas leakage location positioning method belong to the same concept. For the details not described in detail in the technical solution of the computing device, reference can be made to the description of the technical solution of the above gas leakage location positioning method.

[0157] An embodiment of this specification also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the above gas leakage location positioning method.

[0158] The above is a schematic solution of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium and the technical solution of the above gas leakage location positioning method belong to the same concept. For the details not described in detail in the technical solution of the storage medium, reference can be made to the description of the technical solution of the above gas leakage location positioning method.

[0159] An embodiment of this specification also provides a computer program product including a computer program or instructions, which, when executed by a processor, implement the steps of the above gas leakage location positioning method.

[0160] The above is a schematic solution of a computer program product according to this embodiment. It should be noted that the technical solution of this computer program product and the technical solution of the above gas leakage location positioning method belong to the same concept. For the details not described in detail in the technical solution of the computer program product, reference can be made to the description of the technical solution of the above gas leakage location positioning method.

[0161] The above describes specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired result. In certain implementations, multitasking and parallel processing are also possible or may be advantageous.

[0162] The computer instructions include computer program code, which may be in the form of source code, object code, executable files or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of patent practice. For example, in some regions, according to patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0163] It should be noted that for the foregoing method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations. However, those skilled in the art should be aware that the embodiments of this specification are not limited by the described action sequence, because according to the embodiments of this specification, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential for the embodiments of this specification.

[0164] In the above embodiments, the descriptions of each embodiment have their own emphases. For the parts not detailed in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0165] The preferred embodiments of this specification disclosed above are only used to help explain this specification. The alternative embodiments do not describe all the details in detail, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations can be made according to the content of the embodiments of this specification. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the embodiments of this specification, so that those skilled in the art can understand and utilize this specification well.

Claims

1. A method for locating the position of gas leakage, characterized in that, Including: Collecting spectral images corresponding to multiple bands for a gas leakage area by means of a drone, and fusing the spectral images corresponding to the multiple bands to obtain a multispectral image; Extracting the multispectral features of the multispectral image, and constructing a gas concentration heat map corresponding to the gas leakage area according to the multispectral features; Performing gradient calculation on the gas concentration heat map to obtain gradient change information between the gas concentration values and the heat map pixels in the gas concentration heat map, detecting the gas concentration heat map according to the preset sliding window according to the gradient change information, determining multiple candidate heat map pixels that meet the gas leakage detection conditions in the gas concentration heat map according to the detection results, determining the candidate gas concentration values corresponding to each candidate heat map pixel, comparing the candidate gas concentration values with a preset concentration threshold, selecting the candidate heat map pixels corresponding to the candidate gas concentration values greater than the concentration threshold according to the comparison results to construct a pixel set, performing weighted average calculation on the heat map pixels included in the pixel set to obtain the central position coordinates, and determining an abnormal image area in the gas concentration heat map according to the central position coordinates; Constructing a flight path for the drone in the gas leakage area according to the abnormal image area, and driving the drone to locate the gas leakage position based on the flight path.

2. The gas leakage position locating method according to claim 1, characterized in that, The step of collecting spectral images corresponding to multiple bands for a gas leakage area by means of a drone, and fusing the spectral images corresponding to the multiple bands to obtain a multispectral image includes: Collecting images of a gas leakage area at a set angle by means of a multispectral sensor of a drone to obtain spectral images corresponding to multiple bands; Fusing the spectral images corresponding to the multiple bands according to an image registration strategy to obtain a multispectral image.

3. The gas leakage location positioning method according to claim 1, characterized in that The step of extracting the multispectral features of the multispectral image, and constructing a gas concentration heat map corresponding to the gas leakage area according to the multispectral features includes: Processing the multispectral image by using a linear discriminant algorithm to obtain multispectral image features; Inputting the multispectral image features into a regression model for processing to obtain gas concentration value distribution information corresponding to the gas leakage area; Mapping the gas concentration value distribution information to the multispectral image, and generating a gas concentration heat map corresponding to the gas leakage area according to the mapping result.

4. The gas leakage position locating method according to claim 1, wherein The step of constructing a flight path for the drone in the gas leakage area according to the abnormal image area includes: When there are multiple abnormal image areas in the gas concentration heat map, respectively matching each abnormal image area with the flight direction of the drone; Determining a target abnormal image area according to the matching result, and constructing a flight path for the drone in the gas leakage area according to the target abnormal image area.

5. The gas leakage position positioning method according to claim 1, characterized in that, The step of constructing a flight path for the drone in the gas leakage area according to the abnormal image area includes: Calculate the image distance between the abnormal image area and the UAV according to the gradient change information, and compare the image distance with a preset distance threshold; When the image distance is less than the distance threshold, determine the original flight parameters corresponding to the UAV, and construct a flight path for the UAV in the gas leakage area based on the original flight parameters.

6. The gas leakage position locating method according to claim 5, wherein After the step of comparing the image distance with the preset distance threshold is executed, it further includes: When the image distance is greater than or equal to the distance threshold, determine the flight parameters to be adjusted corresponding to the UAV; Adjust the flight parameters to be adjusted according to the gradient change information and the image distance to obtain target flight parameters; Construct a flight path for the UAV in the gas leakage area based on the target flight parameters.

7. The gas leakage position locating method according to any one of claims 1 to 6, characterized in that The driving the UAV to locate the gas leakage position based on the flight path includes: Drive the UAV to move in the gas leakage area based on the flight path; When the position after the UAV moves does not meet the positioning condition, calculate a new sampling position according to the heading angle of the UAV and the position after the UAV moves; Drive the UAV to move to the new sampling position, and execute the step of collecting spectral images corresponding to multiple bands for the gas leakage area by the UAV until the position after the UAV moves meets the positioning condition, and determine the gas leakage position.

8. A gas leakage location positioning device, characterized in that, It includes: A collection module, configured to collect spectral images corresponding to multiple bands for the gas leakage area by the UAV, and fuse the spectral images corresponding to the multiple bands to obtain a multi-spectral image; An extraction module, configured to extract multi-spectral features of the multi-spectral image, and construct a gas concentration heat map corresponding to the gas leakage area according to the multi-spectral features; An identification module, configured to perform gradient calculation on the gas concentration heat map to obtain gradient change information between the gas concentration value and the heat map pixel points in the gas concentration heat map, detect the gas concentration heat map according to the preset sliding window according to the gradient change information, determine multiple candidate heat map pixel points that meet the gas leakage detection conditions in the gas concentration heat map according to the detection results, determine the candidate gas concentration values corresponding to each candidate heat map pixel point, and compare the candidate gas concentration values with a preset concentration threshold, select the candidate heat map pixel points corresponding to the candidate gas concentration values greater than the concentration threshold according to the comparison results to construct a pixel point set, perform weighted average calculation on the heat map pixel points included in the pixel point set to obtain the central position coordinates, and determine the abnormal image area in the gas concentration heat map according to the central position coordinates; A positioning module, configured to construct a flight path for the UAV in the gas leakage area according to the abnormal image area, and drive the UAV to locate the gas leakage position based on the flight path.

9. A computing device, characterized in that, It includes: A memory and a processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that, It stores computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

11. A computer program product, characterized in that, It includes a computer program or instructions. When the computer program or instructions are executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images