Automatic AI picture adjusting method and device for detecting gas scene

Abstract

The invention provides an AI picture automatic adjustment method and device for detecting a gas scene, and the method employs an image AI analysis mode to dynamically adjust the position distribution and coordinates of a leakage point or a point location with a leakage risk in the gas scene, dynamically and adaptively adjust the image output of the leakage point or the point location with the leakage risk, and automatically zoom in or out. The device not only saves manpower, is simple and clear, but also avoids misoperation caused by manual control.

Description

Technical Field

[0001] The present invention relates to the field of gas detection technology, and in particular to an AI picture automatic adjustment method and device for detecting gas scenes. Background Art

[0002] As a clean energy, natural gas is increasingly used by people. However, since natural gas is highly flammable, it has high requirements for gas usage scenarios. The commonly used detection method is generally a simple fixed-position camera that can only detect one direction. In addition, the detection screen is a fixed screen, and it is impossible to automatically zoom in on the location where there may be a gas leak, which is extremely inconvenient during use. Summary of the invention

[0003] The present invention aims to provide an AI picture automatic adjustment method and device for detecting gas scenes that overcomes the above-mentioned problems or at least partially solves the above-mentioned problems.

[0004] To achieve the above object, the technical solution of the present invention is specifically implemented as follows:

[0005] One aspect of the present invention provides an AI picture automatic adjustment method for detecting a gas scene, comprising:

[0006] Collecting the image to be analyzed collected by the gas scene video collector; the image to be analyzed is a fixed large field of view image;

[0007] The image to be analyzed is analyzed by using a gas scene detection algorithm to calculate the location and coordinate information of the leakage point or the point with leakage risk in a large range of video;

[0008] The motion position and zoom position of the pan-tilt screen are calculated according to the corresponding relationship between the image to be analyzed and the pan-tilt, and the pan-tilt screen is used as the output video image, wherein the analyzed leakage point or the point with leakage risk is located within the pan-tilt video range;

[0009] in:

[0010] The image to be analyzed is analyzed by a gas scene detection algorithm to calculate the location and coordinate information of the leakage point or the point with leakage risk in a large range of video, including:

[0011] Determine the center O of the video screen of the video collector;

[0012] According to the coordinates of the center O of the video screen of the video collector, the coordinates of the suspected gas leakage point P1 detected by the gas scene detection algorithm are determined to be (x, y);

[0013] The step of calculating the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt screen, and using the pan-tilt screen as the output video image comprises:

[0014] Align the video collector with a drawing board on which a scale is fixed, wherein the scale is located in the horizontal middle position of the video screen of the video collector;

[0015] Move the drawing board horizontally so that the scale fills the video screen of the video collector, and record the scale values ​​s1 and s2 at the left and right ends of the video collector screen at this time. At this time, the length of the horizontal range of the scale in the monitoring screen is d1=s2-s1;

[0016] The drawing board is translated toward the video collector by a distance d, and the scale values ​​s3 and s4 at the left and right ends of the video screen of the video collector are recorded. At this time, the length of the horizontal range of the scale in the video collector screen is d2=s4-s3;

[0017] Create a mathematical model isosceles triangle and calculate the field of view angle α according to the following formula;

[0018]

[0019] According to the current video's field of view angle α and P1's coordinates (x, y), the horizontal angle Pan required for the PTZ to be centered on P1 and the vertical PTZ rotation angle Tilt are calculated using the following formulas;

[0020]

[0021] Optionally, the image to be analyzed is collected by using a wide-angle lens; or the image to be analyzed is obtained by synchronously collecting video images through multiple lenses, and the video images synchronously collected by the multiple lenses are spliced ​​together.

[0022] Optionally, the step of calculating the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt screen, and using the pan-tilt screen as the output video image includes:

[0023] The video collector including the mechanical pan-tilt head performs image analysis on the image to be analyzed, adjusts the position and zoom position of the pan-tilt head according to the corresponding relationship, and uses the pan-tilt head image as the output video image; or

[0024] The video collector without a mechanical pan / tilt head performs cropping based on the image to be analyzed, and performs cropping as an output video image according to the analyzed leakage point or the coordinate information of the point with leakage risk.

[0025] Another aspect of the present invention provides an AI picture automatic adjustment device for detecting a gas scene, comprising:

[0026] An acquisition module, used to acquire the image to be analyzed acquired by the gas scene video collector; the image to be analyzed is a fixed large field of view image;

[0027] An analysis module, used to analyze the image to be analyzed by using a gas scene detection algorithm, and calculate the location and coordinate information of the leakage point or the point with leakage risk in a large range of video;

[0028] An output module, used to calculate the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt, and use the pan-tilt screen as the output video image, wherein the analyzed leakage point or the point with leakage risk is located within the pan-tilt video range;

[0029] in:

[0030] The analysis module analyzes the image to be analyzed by using a gas scene detection algorithm in the following manner to calculate the location and coordinate information of the leakage point or the point with leakage risk in a large-scale video:

[0031] Determine the center O of the video screen of the video collector;

[0032] According to the coordinates of the center O of the video screen of the video collector, the coordinates of the suspected gas leakage point P1 detected by the gas scene detection algorithm are determined to be (x, y);

[0033] The output module calculates the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt screen in the following manner, and uses the pan-tilt screen as the output video image:

[0034] Align the video collector with a drawing board on which a scale is fixed, wherein the scale is located in the horizontal middle position of the video screen of the video collector;

[0035] Move the drawing board horizontally so that the scale fills the video screen of the video collector, and record the scale values ​​s1 and s2 at the left and right ends of the video collector screen at this time. At this time, the length of the horizontal range of the scale in the monitoring screen is d1=s2-s1;

[0036] The drawing board is translated toward the video collector by a distance d, and the scale values ​​s3 and s4 at the left and right ends of the video screen of the video collector are recorded. At this time, the length of the horizontal range of the scale in the video collector screen is d2=s4-s3;

[0037] Create a mathematical model isosceles triangle and calculate the field of view angle α according to the following formula;

[0038]

[0039] According to the current video's field of view angle α and P1's coordinates (x, y), the horizontal angle Pan required for the PTZ to be centered on P1 and the vertical PTZ rotation angle Tilt are calculated using the following formulas;

[0040]

[0041] Optionally, the image to be analyzed is collected by using a wide-angle lens; or the image to be analyzed is obtained by synchronously collecting video images through multiple lenses, and the video images synchronously collected by the multiple lenses are spliced ​​together.

[0042] Optionally, the output module is used to calculate the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt screen in the following manner, and using the pan-tilt screen as the output video image includes:

[0043] The video collector including the mechanical pan-tilt head performs image analysis on the image to be analyzed, adjusts the position and zoom position of the pan-tilt head according to the corresponding relationship, and uses the pan-tilt head image as the output video image; or

[0044] The video collector without a mechanical pan / tilt head performs cropping based on the image to be analyzed, and performs cropping as an output video image according to the analyzed leakage point or the coordinate information of the point with leakage risk.

[0045] Another aspect of the present invention provides an AI picture automatic adjustment device for detecting gas scenes, comprising: a video collector, a network module, a cloud server, and a video output terminal; wherein:

[0046] The video collector is used to transmit the collected video to the cloud server through the network module carried thereon;

[0047] The cloud server is used to analyze the concentration of combustible gas, smoke, illegal intruders and / or flames in the video collected by the video collector; if there is no abnormality, the normal monitoring video is transmitted; if there is an abnormality, the abnormal point is analyzed; and the graphics obtained after calculation are output through the video output terminal;

[0048] Wherein, the cloud server is specifically used to determine the center O of the video screen of the video collector; according to the coordinates of the center O of the video screen of the video collector, determine the coordinates of the suspected gas leakage point P1 detected by the gas scene detection algorithm as (x, y); align the video collector with a drawing board with a scale fixed on it, wherein the scale is in the horizontal middle position of the video screen of the video collector; horizontally move the drawing board so that the scale fills the video screen of the video collector, and record the scale values ​​s1 and s2 at the left and right ends of the video collector screen at this time. At this time, the length of the horizontal range of the scale in the monitoring screen is d1=s2-s1; translate the drawing board toward the video collector by a distance d, record the scale values ​​s3 and s4 at the left and right ends of the video screen of the video collector at this time, and at this time, the length of the horizontal range of the scale in the video collector screen is d2=s4-s3; create a mathematical model isosceles triangle, and calculate the field of view angle α according to the following formula; α=2*β;According to the field of view angle α of the current video and the coordinates (x, y) of P1, the horizontal angle Pan required for the PTZ to be centered on P1 and the vertical PTZ rotation angle Tilt are calculated by the following formula;

[0049] Optionally, the cloud server is also used to automatically complete the picture when the image is not clear.

[0050] Optionally, the video output terminal includes: a wired or wireless monitor, or a wireless remote handheld device.

[0051] Optionally, a mechanical pan / tilt is provided under the video collector;

[0052] The cloud server is also used to control the movement of the mechanical pan / tilt platform.

[0053] It can be seen that the AI ​​image automatic adjustment method and device for detecting gas scenes provided by the present invention adopt an image AI analysis method, the position distribution and coordinates of leakage points or points with leakage risks in dynamic gas scenes, and dynamically and adaptively adjust the image output of leakage points or points with leakage risks, and automatically zoom in or out, which not only saves manpower, is simple and clear, but also avoids misoperation of human control. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0055] Figure 1 A flowchart of an AI image automatic adjustment method for detecting a gas scene provided by an embodiment of the present invention;

[0056] Figure 2 A front view of a suspected gas leak point in a video provided by an embodiment of the present invention;

[0057] Figure 3 A mathematical model diagram of video field of view angle measurement provided by an embodiment of the present invention;

[0058] Figure 4 A schematic diagram of the gas scene adaptive effect provided by an embodiment of the present invention;

[0059] Figure 5 A schematic diagram of the structure of an AI image automatic adjustment device for detecting gas scenes provided by an embodiment of the present invention;

[0060] Figure 6 A schematic diagram of the structure of another AI image automatic adjustment device for detecting gas scenes provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0061] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.

[0062] Figure 1 A flowchart of an AI picture automatic adjustment method for detecting a gas scene provided by an embodiment of the present invention is shown, see Figure 1 , the AI ​​picture automatic adjustment method for detecting gas scenes provided by the embodiment of the present invention includes:

[0063] S1, collect the image to be analyzed collected by the gas scene video collector; the image to be analyzed is a fixed large field of view image.

[0064] Specifically, the present invention can collect a gas scene video collector image - a fixed large field of view image.

[0065] As an optional implementation of the embodiment of the present invention, the image to be analyzed is collected by using a wide-angle lens; or the image to be analyzed is obtained by synchronously collecting video images through multiple lenses respectively, and the video images synchronously collected by the multiple lenses are spliced.

[0066] Specifically, the present invention can use a wide-angle lens to capture a large-field image of the gas scene as image one, or multiple lenses can synchronously capture video images separately, and then splice the video images synchronously captured by multiple lenses to obtain a large-field image as image one.

[0067] S2, analyze the image to be analyzed through the gas scene detection algorithm, and calculate the location and coordinate information of the leakage point or the point with leakage risk in the large-scale video;

[0068] Among them: the image to be analyzed is analyzed by the gas scene detection algorithm, and the position and coordinate information of the leakage point or the point with leakage risk in the large-scale video are calculated, including: determining the center O of the video screen of the video collector; according to the coordinates of the center O of the video screen of the video collector, determining the coordinates of the suspected gas leakage point P1 detected by the gas scene detection algorithm as (x, y).

[0069] Specifically, the present invention can analyze the image 1 through the gas scene detection algorithm to calculate the position and coordinate information of the leakage point or the point with leakage risk in the large-scale video.

[0070] When implementing it, Figure 2 As shown, the gas scene detection algorithm refers to the "Research and Application of Gas Leakage Monitoring Technology of Gas Cloud Imaging Camera". Through this algorithm, the location and coordinate information of the leakage point or the leakage risk point in the current video screen can be obtained; P1 is the suspected gas leakage point detected by the gas scene detection algorithm, and the coordinates are (x, y); O is the center of the video screen, and the coordinates are (960, 540).

[0071] S3, calculating the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt, and using the pan-tilt screen as the output video image, wherein the analyzed leakage point or the point with leakage risk is located within the pan-tilt video range;

[0072] in:

[0073] According to the corresponding relationship between the image to be analyzed and the PTZ, the motion position and zoom position of the PTZ image are calculated, and the PTZ image is used as the output video image including:

[0074] Align the video collector with a drawing board with a scale fixed on it, wherein the scale is located in the horizontal middle position of the video screen of the video collector;

[0075] Move the drawing board horizontally so that the scale fills the video screen of the video collector, and record the scale values ​​s1 and s2 at the left and right ends of the video collector screen at this time. At this time, the length of the horizontal range of the scale in the monitoring screen is d1 = s2-s1;

[0076] Move the drawing board toward the video collector by a distance d, and record the scale values ​​s3 and s4 at the left and right ends of the video screen of the video collector. At this time, the horizontal length of the scale in the video collector screen is d2 = s4-s3;

[0077] Create a mathematical model isosceles triangle (such as Figure 3 As shown), the field of view angle α is calculated according to the following formula;

[0078]

[0079] According to the current video's field of view angle α and P1's coordinates (x, y), the horizontal angle Pan required for the PTZ to be centered on P1 and the vertical PTZ rotation angle Tilt are calculated using the following formulas;

[0080]

[0081] Specifically, the present invention can calculate the movement position and zoom position of the pan-tilt screen according to the corresponding relationship between image 1 and the mechanical pan-tilt screen, so that the analyzed leakage points or points with leakage risks are just within the pan-tilt video range, and the pan-tilt screen is used as the output video image.

[0082] As an optional implementation manner of an embodiment of the present invention, the movement position and zoom position of the pan-tilt screen are calculated according to the correspondence between the image to be analyzed and the pan-tilt, and the pan-tilt screen is used as the output video image, including: a video collector with a mechanical pan-tilt performs image analysis on the image to be analyzed, adjusts the position and zoom position of the pan-tilt according to the corresponding relationship, and uses the pan-tilt screen as the output video image; or a video collector without a mechanical pan-tilt performs cropping based on the image to be analyzed, and crops according to the analyzed leakage point or the coordinate information of the point with leakage risk as the output video image.

[0083] Specifically, the present invention may include a mechanical pan-tilt head and a zoom camera. A video collector including a pan-tilt head is used to perform image analysis through image 1, and then the position of the pan-tilt head and the zoom position are adjusted according to the corresponding relationship; or a video collector without a mechanical pan-tilt head can be used to perform cropping based on the large-viewing field image 1, and cropping is performed according to the coordinate information of the analyzed leakage point or the point with leakage risk as the output video image.

[0084] in, Figure 4 The adaptive effect of the gas scene is shown, and possible leakage points are given as examples. Through the AI ​​picture automatic adjustment method for detecting gas scenes provided by the embodiment of the present invention, the leakage point or the point with leakage risk in the gas use scene is judged by the AI ​​algorithm, and then the computer AI automatically zooms in on the point, which can achieve dynamic real-time adjustment.

[0085] It can be seen that the AI ​​image automatic adjustment method for detecting gas scenes provided by the present invention adopts an image AI analysis method, the position distribution and coordinates of leakage points or points with leakage risks in dynamic gas scenes, and dynamically and adaptively adjusts the image output of leakage points or points with leakage risks, automatically zooming in or out, which not only saves manpower, is simple and clear, but also avoids misoperation of human control.

[0086] Figure 5 The structure diagram of an AI picture automatic adjustment device for detecting a gas scene provided by an embodiment of the present invention is shown. The AI ​​picture automatic adjustment device for detecting a gas scene applies the above method. The following is only a brief description of the structure of the AI ​​picture automatic adjustment device for detecting a gas scene. For other matters not covered, please refer to the relevant description in the above AI picture automatic adjustment method for detecting a gas scene. Figure 5 , an AI picture automatic adjustment device for detecting a gas scene provided by an embodiment of the present invention comprises:

[0087] The acquisition module is used to acquire the image to be analyzed acquired by the gas scene video collector; the image to be analyzed is a fixed large field of view image;

[0088] An analysis module is used to analyze the image to be analyzed through a gas scene detection algorithm, and calculate the location and coordinate information of the leakage point or the point with leakage risk in a large range of video;

[0089] The output module is used to calculate the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt, and use the pan-tilt screen as the output video image, wherein the analyzed leakage point or the point with leakage risk is located within the pan-tilt video range; wherein:

[0090] The analysis module uses the gas scene detection algorithm to analyze the image to be analyzed in the following way to calculate the location and coordinate information of the leakage point or the point with leakage risk in the large-scale video:

[0091] Determine the center O of the video screen of the video collector;

[0092] According to the coordinates of the center O of the video screen of the video collector, the coordinates of the suspected gas leakage point P1 detected by the gas scene detection algorithm are determined to be (x, y);

[0093] The output module calculates the motion position and zoom position of the PTZ image according to the corresponding relationship between the image to be analyzed and the PTZ in the following way, and uses the PTZ image as the output video image:

[0094] Align the video collector with a drawing board with a scale fixed on it, wherein the scale is located in the horizontal middle position of the video screen of the video collector;

[0095] Move the drawing board horizontally so that the scale fills the video screen of the video collector, and record the scale values ​​s1 and s2 at the left and right ends of the video collector screen at this time. At this time, the length of the horizontal range of the scale in the monitoring screen is d1 = s2-s1;

[0096] Move the drawing board toward the video collector by a distance d, and record the scale values ​​s3 and s4 at the left and right ends of the video screen of the video collector. At this time, the horizontal length of the scale in the video collector screen is d2 = s4-s3;

[0097] Create a mathematical model isosceles triangle and calculate the field of view angle α according to the following formula;

[0098]

[0099] According to the current video's field of view angle α and P1's coordinates (x, y), the horizontal angle Pan required for the PTZ to be centered on P1 and the vertical PTZ rotation angle Tilt are calculated using the following formulas;

[0100]

[0101] As an optional implementation of the embodiment of the present invention, the image to be analyzed is collected by using a wide-angle lens; or the image to be analyzed is obtained by synchronously collecting video images through multiple lenses respectively, and the video images synchronously collected by the multiple lenses are spliced.

[0102] As an optional implementation manner of an embodiment of the present invention, the output module is used to calculate the movement position and zoom position of the pan-tilt screen according to the correspondence between the image to be analyzed and the pan-tilt, and use the pan-tilt screen as the output video image, including: a video collector with a mechanical pan-tilt performs image analysis on the image to be analyzed, adjusts the position and zoom position of the pan-tilt according to the corresponding relationship, and uses the pan-tilt screen as the output video image; or a video collector without a mechanical pan-tilt performs cropping based on the image to be analyzed, and crops according to the analyzed leakage point or the coordinate information of the point with leakage risk as the output video image.

[0103] It can be seen that the AI ​​image automatic adjustment device for detecting gas scenes provided by the present invention adopts an image AI analysis method, the position distribution and coordinates of the leakage points or points with leakage risks in the dynamic gas scene, and dynamically and adaptively adjusts the image output of the leakage points or points with leakage risks, and automatically zooms in or out, which not only saves manpower, is simple and clear, but also avoids misoperation of human control.

[0104] Figure 6 A schematic diagram of the structure of another AI picture automatic adjustment device for detecting a gas scene provided by an embodiment of the present invention is shown. Another AI picture automatic adjustment device for detecting a gas scene provided by an embodiment of the present invention includes: a video collector, a network module, a cloud server, and a video output terminal; wherein:

[0105] The video collector is used to transmit the collected video to the cloud server through the network module;

[0106] The cloud server is used to analyze the flammable gas concentration, smoke, illegal intruders and / or flames in the video collected by the video collector; if there is no abnormality, the normal monitoring video is transmitted; if there is an abnormality, the abnormal point is analyzed; and the calculated graphics are output through the video output terminal;

[0107] Among them, the cloud server is specifically used to determine the center O of the video screen of the video collector; according to the coordinates of the center O of the video screen of the video collector, the coordinates of the suspected gas leakage point P1 detected by the gas scene detection algorithm are determined to be (x, y); the video collector is aligned with a drawing board with a scale fixed on it, wherein the scale is in the horizontal middle position of the video screen of the video collector; the drawing board is moved horizontally so that the scale fills the video screen of the video collector, and the scale values ​​s1 and s2 at the left and right ends of the video collector screen are recorded at this time. At this time, the scale is in the monitoring screen, and the length of the horizontal range is d1=s2-s1; the drawing board is translated toward the video collector by a distance d, and the scale values ​​s3 and s4 at the left and right ends of the video screen of the video collector are recorded at this time. At this time, the length of the horizontal range of the scale in the video collector screen is d2=s4-s3; a mathematical model isosceles triangle is created, and the field of view angle α is calculated according to the following formula; α=2*β;According to the field of view angle α of the current video and the coordinates (x, y) of P1, the horizontal angle Pan required for the PTZ to be centered on P1 and the vertical PTZ rotation angle Tilt are calculated by the following formula;

[0108] As an optional implementation of the embodiment of the present invention, the cloud server is also used to automatically complete the picture when the image is not clear.

[0109] As an optional implementation of the embodiment of the present invention, the video output terminal includes: a wired or wireless monitor, or a wireless remote handheld device.

[0110] As an optional implementation of the embodiment of the present invention, a mechanical pan-tilt head is arranged under the video collector; the cloud server is also used to control the movement of the mechanical pan-tilt head.

[0111] Specifically, after the video collector collects the video, it transmits it to the cloud server through the network module. The cloud server is equipped with an artificial intelligence AI program, which can automatically analyze the flammable gas concentration, smoke, illegal intrusion, flame, etc. in the video collected by the video collector. If there is no abnormality, the normal monitoring video is transmitted. If an abnormality occurs, the abnormal point (leakage point or point with leakage risk, including flammable gas concentration, smoke, illegal intrusion, flame, etc.) is analyzed and automatically focused and enlarged. If the image is not clear, the internal artificial intelligence AI program is used to automatically complete the picture. The cloud server outputs the calculated graphics through the video output terminal. The video output terminal can be a wired or wireless monitor, or a wireless remote handheld device. A mechanical pan-tilt can also be installed under the video collector, and the cloud server can control the movement of the mechanical pan-tilt to achieve the effect of increasing the acquisition range of the video collector.

[0112] It can be seen that the AI ​​image automatic adjustment device for detecting gas scenes provided by the present invention adopts the image AI analysis method, the position distribution and coordinates of the leakage points or points with leakage risks in the dynamic gas scene, and the image output of the leakage points or points with leakage risks is dynamically and adaptively adjusted, and automatically enlarged or reduced, which not only saves manpower, is simple and clear, but also avoids the misoperation of human control. Furthermore, the present invention can detect the concentration of combustible gas, smoke, illegal intruders, flames, etc. in the gas scene.

[0113] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. An AI image automatic adjustment method for detecting gas scenes, characterized in that: include: Collecting the image to be analyzed collected by the gas scene video collector; the image to be analyzed is a fixed large field of view image; The image to be analyzed is analyzed by using a gas scene detection algorithm to calculate the location and coordinate information of the leakage point or the point with leakage risk in a large range of video; The motion position and zoom position of the pan-tilt screen are calculated according to the corresponding relationship between the image to be analyzed and the pan-tilt, and the pan-tilt screen is used as the output video image, wherein the analyzed leakage point or the point with leakage risk is located within the pan-tilt video range; in: The image to be analyzed is analyzed by a gas scene detection algorithm to calculate the location and coordinate information of the leakage point or the point with leakage risk in a large range of video, including: Determine the center O of the video screen of the video collector; According to the coordinates of the center O of the video screen of the video collector, the coordinates of the suspected gas leakage point P1 detected by the gas scene detection algorithm are determined to be (x, y); The step of calculating the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt screen, and using the pan-tilt screen as the output video image comprises: Align the video collector with a drawing board on which a scale is fixed, wherein the scale is located in the horizontal middle position of the video screen of the video collector; Move the drawing board horizontally so that the scale fills the video screen of the video collector, and record the scale values ​​s1 and s2 at the left and right ends of the video collector screen at this time. At this time, the length of the horizontal range of the scale in the monitoring screen is d1=s2-s1; The drawing board is translated toward the video collector by a distance d, and the scale values ​​s3 and s4 at the left and right ends of the video screen of the video collector are recorded. At this time, the length of the horizontal range of the scale in the video collector screen is d2=s4-s3; Create a mathematical model isosceles triangle and calculate the field of view angle α according to the following formula; According to the current video's field of view angle α and P1's coordinates (x, y), the horizontal angle Pan required for the PTZ to be centered on P1 and the vertical PTZ rotation angle Tilt are calculated using the following formulas; 2. The method according to claim 1, characterized in that The image to be analyzed is collected by using a wide-angle lens; or the image to be analyzed is collected by using multiple lenses to synchronously collect video images, and the video images synchronously collected by the multiple lenses are spliced ​​to obtain the image.

3. The method according to claim 2, characterized in that The step of calculating the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt screen, and using the pan-tilt screen as the output video image comprises: The video collector including the mechanical pan-tilt head performs image analysis on the image to be analyzed, adjusts the position and zoom position of the pan-tilt head according to the corresponding relationship, and uses the pan-tilt head image as the output video image; or The video collector without a mechanical pan / tilt head performs cropping based on the image to be analyzed, and performs cropping as an output video image according to the analyzed leakage point or the coordinate information of the point with leakage risk.

4. An AI image automatic adjustment device for detecting gas scenes, characterized in that: include: An acquisition module, used to acquire the image to be analyzed acquired by the gas scene video collector; the image to be analyzed is a fixed large field of view image; An analysis module, used to analyze the image to be analyzed by using a gas scene detection algorithm, and calculate the location and coordinate information of the leakage point or the point with leakage risk in a large range of video; An output module, used to calculate the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt, and use the pan-tilt screen as the output video image, wherein the analyzed leakage point or the point with leakage risk is located within the pan-tilt video range; in: The analysis module analyzes the image to be analyzed by using a gas scene detection algorithm in the following manner to calculate the location and coordinate information of the leakage point or the point with leakage risk in a large-scale video: Determine the center O of the video screen of the video collector; According to the coordinates of the center O of the video screen of the video collector, the coordinates of the suspected gas leakage point P1 detected by the gas scene detection algorithm are determined to be (x, y); The output module calculates the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt screen in the following manner, and uses the pan-tilt screen as the output video image: Align the video collector with a drawing board on which a scale is fixed, wherein the scale is located in the horizontal middle position of the video screen of the video collector; Move the drawing board horizontally so that the scale fills the video screen of the video collector, and record the scale values ​​s1 and s2 at the left and right ends of the video collector screen at this time. At this time, the length of the horizontal range of the scale in the monitoring screen is d1=s2-s1; The drawing board is translated toward the video collector by a distance d, and the scale values ​​s3 and s4 at the left and right ends of the video screen of the video collector are recorded. At this time, the length of the horizontal range of the scale in the video collector screen is d2=s4-s3; Create a mathematical model isosceles triangle and calculate the field of view angle α according to the following formula; α＝2*β； According to the current video's field of view angle α and P1's coordinates (x, y), the horizontal angle Pan required for the PTZ to be centered on P1 and the vertical PTZ rotation angle Tilt are calculated using the following formulas; 5. The device according to claim 4, characterized in that The image to be analyzed is collected by using a wide-angle lens; or the image to be analyzed is collected by using multiple lenses to synchronously collect video images, and the video images synchronously collected by the multiple lenses are spliced ​​to obtain the image.

6. The method according to claim 5, characterized in that The output module is used to calculate the motion position and zoom position of the pan-tilt screen according to the corresponding relationship between the image to be analyzed and the pan-tilt screen in the following manner, and the pan-tilt screen is used as the output video image, including: The video collector including the mechanical pan-tilt head performs image analysis on the image to be analyzed, adjusts the position and zoom position of the pan-tilt head according to the corresponding relationship, and uses the pan-tilt head image as the output video image; or The video collector without a mechanical pan / tilt head performs cropping based on the image to be analyzed, and performs cropping as an output video image according to the analyzed leakage point or the coordinate information of the point with leakage risk.

7. An AI image automatic adjustment device for detecting gas scenes, characterized in that: include: Video collector, network module, cloud server, video output terminal; among which: The video collector is used to transmit the collected video to the cloud server through the network module carried thereon; The cloud server is used to analyze the concentration of combustible gas, smoke, illegal intruders and / or flames in the video collected by the video collector; if there is no abnormality, the normal monitoring video is transmitted; if there is an abnormality, the abnormal point is analyzed; and the graphics obtained after calculation are output through the video output terminal; Wherein, the cloud server is specifically used to determine the center O of the video screen of the video collector; according to the coordinates of the center O of the video screen of the video collector, determine the coordinates of the suspected gas leakage point P1 detected by the gas scene detection algorithm as (x, y); align the video collector with a drawing board with a scale fixed on it, wherein the scale is in the horizontal middle position of the video screen of the video collector; horizontally move the drawing board so that the scale fills the video screen of the video collector, and record the scale values ​​s1 and s2 at the left and right ends of the video collector screen at this time. At this time, the length of the horizontal range of the scale in the monitoring screen is d1=s2-s1; translate the drawing board toward the video collector by a distance d, record the scale values ​​s3 and s4 at the left and right ends of the video screen of the video collector at this time, and at this time, the length of the horizontal range of the scale in the video collector screen is d2=s4-s3; create a mathematical model isosceles triangle, and calculate the field of view angle α according to the following formula; α=2*β;According to the field of view angle α of the current video and the coordinates (x, y) of P1, the horizontal angle Pan required for the PTZ to be centered on P1 and the vertical PTZ rotation angle Tilt are calculated by the following formula; 8. The device according to claim 7, characterized in that The cloud server is also used to automatically make the picture clear when the image is not clear.

9. The device according to claim 8, characterized in that The video output terminal includes: a wired or wireless monitor, or a wireless remote handheld device.

10. The device according to claim 9, characterized in that A mechanical pan / tilt platform is arranged under the video collector; The cloud server is also used to control the movement of the mechanical pan / tilt platform.

Images