Artificial intelligence nitrogen dioxide gas detecting and reporting system and operating method thereof

Abstract

An artificial intelligence-based nitrogen dioxide gas detection and reporting system is disclosed for detecting yellow nitrogen dioxide gas emitted from a flue outlet. The system includes an image collector and a computational analysis server. The image collector's monitor lens is positioned at the flue outlet, and the monitor's main unit acquires real-time smoke images generated by the monitor lens. The computational analysis server receives the real-time smoke images and has an artificial intelligence image analysis module. The artificial intelligence image analysis module detects the real-time smoke images using an object detection model to generate analysis information and tracks the analysis information using a trajectory tracking model to generate smoke identification results. Furthermore, the artificial intelligence image analysis module analyzes the smoke identification results using the image analysis model to generate nitrogen dioxide gas concentration information.

Description

Technical Field

[0001] This disclosure relates to a detection system and its operating method, and more particularly to an artificial intelligence-based system for detecting and reporting nitrogen dioxide gas and its operating method, which improves the efficiency of traditional smoke detection operations and reduces overall equipment maintenance costs by using smoke image recognition technology through an image collector and a computing and analysis server. Prior Technology

[0002] Generally, nitric acid solution is used as a reagent in the metal waste treatment and recycling process to dissolve metal compounds. Specifically, the metal is oxidized into metal ions that enter the nitric acid solution, while nitrate ions are reduced to generate gaseous nitrogen dioxide. Nitrogen dioxide treatment systems are as disclosed in Taiwanese Utility Model Patent TW M661682 U (hereinafter referred to as Document 1). However, Document 1 only describes how to treat nitrogen dioxide gas to enable its emission and does not disclose the technology for image recognition of nitrogen dioxide gas. When gaseous nitrogen dioxide is emitted into the air, it forms yellow smoke, commonly known as yellow smoke. In the traditional field of smoke detection, smoke detectors are installed at the flue gas collection point to analyze changes in the concentration of nitrogen dioxide gas in the air. However, installing smoke detectors at the flue gas collection point exposes the sensors to a harsh environment (containing acidic and / or alkaline gases) for extended periods. Acidic environments easily damage sensor equipment, thus requiring frequent sensor replacements, reducing the efficiency of smoke detection operations and increasing maintenance costs. Summary of the Invention

[0003] One of the technical embodiments disclosed herein is an artificial intelligence-based system for detecting and reporting nitrogen dioxide gas. This system can replace the traditional identification technology of installing physical smoke detectors at the point of flue gas convergence with smoke image recognition. This improves the situation where the sensor equipment is easily damaged and needs to be frequently replaced due to the long-term exposure of smoke detectors to acidic gas environment. It can improve the operational efficiency of smoke detection and reduce the overall maintenance cost of equipment.

[0004] This disclosed embodiment provides an AI-based system for detecting and reporting nitrogen dioxide gas. This system detects yellow nitrogen dioxide gas emitted from a flue outlet and includes an image collector and a computational analysis server. The image collector has a monitor lens and a monitor host electrically connected to the monitor lens. The monitor lens is positioned on one side of the flue outlet and is used to capture images of the flue outlet. The monitor host is used to acquire real-time smoke images generated by the monitor lens capturing images of the flue outlet. The computational analysis server is communicatively connected to the image collector and is used to receive the real-time smoke images transmitted from the monitor host of the image collector. The computational analysis server includes an AI image analysis module. This module detects the real-time smoke images based on an object detection model to generate analysis information and performs trajectory tracking on the analysis information based on a trajectory tracking model to generate smoke identification results. The AI ​​image analysis module further analyzes the smoke identification results based on the image analysis model to generate nitrogen dioxide gas concentration information.

[0005] According to one embodiment of this disclosure, the aforementioned computational analysis server further includes an image correction module. The image correction module is communicatively connected to an image collector and is used to perform image correction on real-time smoke exhaust images to obtain a plurality of preprocessing information.

[0006] According to one embodiment of this disclosure, the aforementioned computational analysis server further includes a feature extraction module and a storage module. The feature extraction module is communicatively connected to the image correction module and is used to receive a plurality of preprocessing information and extract a plurality of feature information from the plurality of preprocessing information. The storage module is communicatively connected to the feature extraction module and is used to store the plurality of feature information.

[0007] According to one embodiment of this disclosure, the object detection model is established based on the YOLO algorithm. The artificial intelligence image analysis module is further used to detect multiple preprocessing information of the real-time smoke exhaust image after image correction based on the object detection model in order to generate analysis information. The trajectory tracking model used by the artificial intelligence image analysis module of the computing and analysis server is a two-stage region-based convolutional neural network (R-CNN).

[0008] According to one embodiment of this disclosure, the aforementioned artificial intelligence-based nitrogen dioxide gas detection and reporting system further includes an application terminal device. The application terminal device is communicatively connected to a computing and analysis server and is used to receive smoke identification results and nitrogen dioxide gas concentration information transmitted by the computing and analysis server. The application terminal device can be a personal computer, laptop, tablet computer, or mobile device.

[0009] According to one embodiment of this disclosure, the aforementioned computational analysis server is further used to set an identification area on the real-time smoke exhaust image, and the artificial intelligence image analysis module is further used to detect the identification area according to the object detection model to generate analysis information.

[0010] Another embodiment disclosed herein is an operation method for detecting and reporting nitrogen dioxide gas using artificial intelligence. It can replace the traditional identification technology of installing physical smoke detectors at the smoke collection point with smoke image recognition. This improves the situation where the sensor equipment is easily damaged and needs to be frequently replaced due to the long-term exposure of the smoke detector to the acid gas environment. It can improve the operational efficiency of smoke detection and reduce the overall maintenance cost of the equipment.

[0011] The disclosed embodiments provide an operation method for an artificial intelligence-based nitrogen dioxide gas detection and reporting system to detect yellow nitrogen dioxide gas emitted from a flue outlet. The method includes: capturing images of the flue outlet using a monitor lens of an image collector, enabling the monitor host of the image collector to acquire real-time smoke emission images generated from the captured images, wherein the monitor lens is electrically connected to the monitor host and is positioned on one side of the flue outlet; receiving the real-time smoke emission images transmitted from the monitor host of the image collector via a computational analysis server; performing analysis on the real-time smoke emission images using an artificial intelligence image analysis module of the computational analysis server based on an object detection model to generate analysis information; performing trajectory tracking on the analysis information using the artificial intelligence image analysis module of the computational analysis server based on a trajectory tracking model to generate smoke identification results; and analyzing the smoke identification results using the image analysis model to generate nitrogen dioxide gas concentration information.

[0012] According to one embodiment of this disclosure, receiving real-time smoke exhaust images via a computational analysis server further includes receiving real-time smoke exhaust images via an image correction module of the computational analysis server and performing image correction on the real-time smoke exhaust images to obtain a plurality of preprocessing information.

[0013] According to one embodiment of this disclosure, the above-mentioned operation method of the artificial intelligence-based nitrogen dioxide gas detection and notification system further includes: receiving a plurality of preprocessing information by means of a feature extraction module of a computing analysis server, and extracting a plurality of feature information from the plurality of preprocessing information; and storing the plurality of feature information by means of a storage module of a computing analysis server.

[0014] According to one embodiment of this disclosure, the above-mentioned operation method of the artificial intelligence-based nitrogen dioxide gas detection and reporting system further includes receiving smoke identification results and nitrogen dioxide gas concentration information via an application terminal device connected to a computing and analysis server, wherein the application terminal device is a personal computer, laptop, tablet computer, or mobile device.

[0015] In summary, this disclosure utilizes an image collector within an AI-powered nitrogen dioxide gas detection and reporting system to acquire real-time smoke images generated at the flue outlet. The AI ​​image analysis module on the AI-powered nitrogen dioxide gas detection and reporting system's computational analysis server can detect the real-time smoke images using an object detection model to generate analytical information. Next, the AI ​​image analysis module can track the analyzed information using a trajectory tracking model to generate smoke identification results. Specifically, the AI ​​image analysis module first determines whether there are smoke-like objects in the real-time smoke images, and then uses a trajectory tracking model to determine whether these smoke-like objects are indeed smoke. After confirming it as smoke, the AI ​​image analysis module analyzes the smoke identification results using the image analysis model to generate nitrogen dioxide gas concentration information. By using artificial intelligence to detect nitrogen dioxide gas and identify smoke images in the reporting system, the traditional identification technology of installing physical smoke detectors at the smoke duct convergence point can be replaced. This can improve the situation where the sensor equipment is easily damaged and needs to be frequently replaced due to the long-term exposure of smoke detectors to acidic gas environment. It can improve the efficiency of smoke detection and reduce the overall maintenance cost of equipment. Simple Explanation of the Diagram

[0016] One embodiment of this disclosure is best understood when read in conjunction with the accompanying figures, from the following detailed description. It should be emphasized that, according to standard industrial practice, the various features are not drawn to scale and are for illustrative purposes only. In fact, the dimensions of the various features may be arbitrarily increased or decreased for clarity of explanation. Figure 1 illustrates a block diagram of an artificial intelligence-based system for detecting and reporting nitrogen dioxide gas according to an embodiment of this disclosure. Figure 2 illustrates a block diagram of an image collector according to one embodiment of the present disclosure. Figure 3 illustrates a block diagram of a computational analysis server according to one embodiment of the present disclosure. Figures 4 and 5 illustrate schematic diagrams of real-time smoke exhaust images captured by a monitor lens at different stages of a flue outlet according to some embodiments of this disclosure. Figure 6 illustrates a flowchart of the operation method of a computational analysis server according to an embodiment of this disclosure. Figure 7 illustrates a flowchart of an operation method for a nitrogen dioxide gas detection and reporting system using artificial intelligence according to an embodiment of this disclosure. Implementation

[0017] The following disclosure of embodiments provides many different implementations, or examples, for carrying out different features of the provided object. Specific examples of elements and arrangements are described below to simplify the subject matter. Of course, these examples are merely illustrative and are not intended to be limiting. Furthermore, element symbols and / or letters may be repeated in various examples. This repetition is for simplicity and clarity and does not in itself specify the relationship between the various embodiments and / or configurations discussed.

[0018] Spatial relative terms such as “below,” “under,” “lower,” “above,” and “upper” are used herein for descriptive purposes to describe the relationship between one element or feature and another, as shown in the accompanying drawings. Spatial relative terms are intended to cover different orientations of the device in use or operation other than those shown in the accompanying drawings. The device may be oriented in other ways (rotated 90 degrees or otherwise), and the spatial relative descriptors used herein shall be interpreted accordingly.

[0019] Please refer to Figures 1 to 3. Figure 1 shows a block diagram of an AI-based nitrogen dioxide gas detection and notification system 100 according to an embodiment of this disclosure. Figure 2 shows a block diagram of an image collector 110 according to an embodiment of this disclosure. Figure 3 shows a block diagram of a computing and analysis server 120 according to an embodiment of this disclosure. In Figures 1 to 3, the AI-based nitrogen dioxide gas detection and notification system 100 includes an image collector 110 and a computing and analysis server 120. The image collector 110 has a monitor lens 112 and a monitor host 114 electrically connected to the monitor lens 112. The computing and analysis server 120 is communicatively connected to the image collector 110. The computing and analysis server 120 has an AI image analysis module 122. The AI-based nitrogen dioxide gas detection and notification system 100 further includes an application terminal device 130. The application terminal device 130 is connected to the computing and analysis server 120. The application terminal device 130 can be a personal computer, a laptop, a tablet computer, or a mobile device.

[0020] Please refer to Figures 4 to 6. Figures 4 and 5 illustrate schematic diagrams of real-time smoke exhaust images C captured by a monitor lens 112 at different stages of flue gas outlet A according to some embodiments of this disclosure. Figure 6 illustrates a flowchart of the operation method of a computational analysis server 120 according to an embodiment of this disclosure. In Figures 1 to 6, an artificial intelligence nitrogen dioxide gas detection and reporting system 100 is used to detect the yellow nitrogen dioxide gas B emitted from flue gas outlet A. The monitor lens 112 can be installed on one side of flue gas outlet A and is used to capture images of flue gas outlet A, while the monitor host 114 is used to acquire the real-time smoke exhaust images C generated by the monitor lens 112 capturing images of flue gas outlet A.

[0021] In some embodiments, the computational analysis server 120 is used to receive real-time smoke exhaust images C transmitted by the monitor host 114 of the image collector 110. The computational analysis server 120 is further used to set an identification area D on the real-time smoke exhaust images C, and the identification area D may be located at the outlet of the flue A. In other embodiments, physical smoke detectors may be installed around the flue A for assistance and comparison to improve the detection effect of the artificial intelligence-based nitrogen dioxide gas detection and notification system 100.

[0022] Furthermore, the operation method of the computational analysis server 120 may include the following steps. First, in step S11, image correction is performed. The computational analysis server 120 further includes an image correction module 124. The image correction module 124 is communicatively connected to the image collector 110 and is used to perform image correction on the real-time smoke exhaust image C to obtain a plurality of preprocessing information. For example, the image correction module 124 is used to perform preprocessing steps such as brightness adjustment, angle adjustment, noise removal, image blurring, image enhancement, or image scaling on the real-time smoke exhaust image C.

[0023] Next, in step S12, object detection is performed. The AI ​​image analysis module 122 of the computational analysis server 120 is used to detect the real-time smoke exhaust image C according to the object detection model to generate analysis information, or the AI ​​image analysis module 122 is used to detect the identification area D in the real-time smoke exhaust image C according to the object detection model to generate analysis information, or the AI ​​image analysis module 122 is used to detect multiple pre-processed information of the real-time smoke exhaust image C after image correction according to the object detection model to generate analysis information. In detail, the AI ​​image analysis module 122 of the computational analysis server 120 can first determine whether the identification area D in the real-time smoke exhaust image C has an object resembling smoke.

[0024] Next, in step S13, trajectory tracking can be performed. The artificial intelligence image analysis module 122 of the computational analysis server 120 can perform trajectory tracking on the analyzed information according to the trajectory tracking model to generate smoke identification results. In detail, the artificial intelligence image analysis module 122 of the computational analysis server 120 uses the trajectory tracking model to determine whether objects resembling smoke in multiple real-time smoke exhaust images C move upwards like smoke to confirm whether it is indeed nitrogen dioxide yellow gas B.

[0025] Next, in step S14, smoke concentration analysis can be performed. After determining, through the trajectory tracking model, that the object in the real-time smoke image C is indeed nitrogen dioxide yellow gas B, the AI ​​image analysis module 122 of the computational analysis server 120 analyzes the smoke identification results according to the image analysis model to generate nitrogen dioxide gas concentration information. For example, the AI ​​image analysis module 122 of the computational analysis server 120 can identify the smoke concentration and smoke color of nitrogen dioxide yellow gas B in the real-time smoke image C. In some embodiments, the object detection model used by the AI ​​image analysis module 122 of the computational analysis server 120 is established based on the YOLO algorithm, and the trajectory tracking model used by the AI ​​image analysis module 122 of the computational analysis server 120 is a two-stage region-based convolutional neural network (R-CNN).

[0026] During the training phase of the AI ​​image analysis module 122 in the computational analysis server 120, the computational analysis server 120 further includes a feature extraction module 126 and a storage module 128. The feature extraction module 126 is communicatively connected to the image correction module 124 and is used to receive a plurality of preprocessing information and extract a plurality of feature information from the plurality of preprocessing information. The storage module 128 is communicatively connected to the feature extraction module 126 and is used to store the plurality of feature information. This feature information can help the AI ​​image analysis module 122 determine whether the smoke-like object in the real-time smoke exhaust image C is nitrogen dioxide yellow gas B during the testing phase, and the trained data and model can be stored in the storage module 128.

[0027] In some implementations, the application terminal device 130 is used to receive smoke identification results and nitrogen dioxide gas concentration information transmitted by the computing and analysis server 120. For example, the smoke identification results and nitrogen dioxide gas concentration information transmitted by the computing and analysis server 120 can be sent to the mailbox in the application terminal device 130 via email. The email can include information about the shooting location, shooting time, smoke identification status, and on-site photos. Alternatively, the smoke identification results and nitrogen dioxide gas concentration information transmitted by the computing and analysis server 120 can be sent to the application program of the application terminal device 130 via official communication software. The message can include information about the shooting location, shooting time, smoke identification status, and on-site photos.

[0028] The following description will explain the operation of the artificial intelligence-based nitrogen dioxide gas detection and reporting system. The previously described component connections, materials, and functions will not be repeated here; they will be stated in the preceding text.

[0029] Please refer to Figure 7, which illustrates a flowchart of an operation method for an artificial intelligence-based nitrogen dioxide gas detection and notification system according to an embodiment of this disclosure. The operation method of the artificial intelligence-based nitrogen dioxide gas detection and notification system is used to detect yellow nitrogen dioxide gas emitted from a flue. The operation method includes the following steps: First, in step S21, the flue is photographed by a monitor lens of an image collector, allowing the monitor host of the image collector to acquire a real-time smoke emission image generated by photographing the flue. The monitor lens is electrically connected to the monitor host and is positioned on one side of the flue. Next, in step S22, the real-time smoke emission image transmitted by the monitor host of the image collector is received by a computational analysis server. Next, in step S23, the artificial intelligence image analysis module of the computational analysis server detects the real-time smoke emission image based on an object detection model to generate analysis information. Next, in step S24, the artificial intelligence image analysis module of the computational analysis server performs trajectory tracking on the analysis information based on a trajectory tracking model to generate smoke identification results. Next, in step S25, the AI ​​image analysis module of the computational analysis server analyzes the smoke identification results according to the image analysis model to generate nitrogen dioxide gas concentration information. The above steps will be described in detail below.

[0030] In Figures 1 to 7, the monitor lens 112 of the image collector 110 first captures an image of the flue outlet A, allowing the monitor host 114 of the image collector 110 to acquire a real-time smoke exhaust image C generated from the captured image of the flue outlet A. The monitor lens 112 is electrically connected to the monitor host 114 and is positioned on one side of the flue outlet A. Next, the real-time smoke exhaust image C transmitted from the monitor host 114 of the image collector 110 is received by the computing and analysis server 120. Then, the artificial intelligence image analysis module 122 of the computing and analysis server 120 detects the real-time smoke exhaust image C according to an object detection model to generate analysis information. Finally, the artificial intelligence image analysis module 122 of the computing and analysis server 120 performs trajectory tracking on the analysis information according to a trajectory tracking model to generate smoke identification results. Next, the AI ​​image analysis module 122 of the computing and analysis server 120 can analyze the smoke identification results according to the image analysis model to generate nitrogen dioxide gas concentration information.

[0031] In some embodiments, receiving the real-time smoke exhaust image C via the computing and analysis server 120 further includes receiving the real-time smoke exhaust image C via the image correction module 124 of the computing and analysis server 120, and performing image correction on the real-time smoke exhaust image C to obtain a plurality of preprocessing information. Furthermore, the feature extraction module 126 of the computing and analysis server 120 can also receive the plurality of preprocessing information, extract a plurality of feature information from the plurality of preprocessing information, and store the plurality of feature information via the storage module 128 of the computing and analysis server 120. In some embodiments, the smoke identification results and nitrogen dioxide gas concentration information can be received via an application terminal device 130 connected to the computing and analysis server 120, wherein the application terminal device 130 is a personal computer, laptop, tablet computer, or mobile device.

[0032] In summary, the image collector 110 of the AI-powered nitrogen dioxide gas detection and reporting system 100 can acquire real-time smoke emission images C generated from the flue outlet A. The AI-powered image analysis module 122 of the AI-powered nitrogen dioxide gas detection and reporting system 100's computing and analysis server 120 can detect the real-time smoke emission images C using an object detection model to generate analysis information. Next, the AI-powered image analysis module 122 can track the analysis information using a trajectory tracking model to generate smoke identification results. Specifically, the AI-powered image analysis module 122 first determines whether there are smoke-like objects in the real-time smoke emission images C, and then uses a trajectory tracking model to determine whether the smoke-like objects are indeed nitrogen dioxide yellow gas B. After confirming that it is nitrogen dioxide yellow gas B, the AI-powered image analysis module 122 can analyze the smoke identification results using the image analysis model to generate nitrogen dioxide gas concentration information. The use of the AI-based nitrogen dioxide gas detection and notification system 100 is not affected by acidic or alkaline environments. Furthermore, the smoke image recognition of the AI-based nitrogen dioxide gas detection and notification system 100 can replace the traditional identification technology that involves installing physical smoke detectors at the point of convergence in the flue (such as flue outlet A). This can improve the situation where the sensor equipment is easily damaged and requires frequent replacement of smoke detector equipment due to long-term exposure to acidic gas environments. It can improve the operational efficiency of smoke detection and reduce the overall maintenance cost of equipment.

[0033] The foregoing outlines the features of several embodiments, enabling those skilled in the art to better understand the nature of this disclosure. Those skilled in the art should understand that they can readily use this disclosure as the basis for designing or modifying other processes and structures to achieve the same objectives and / or advantages as the embodiments described herein. Those skilled in the art should also recognize that such equivalent constructions do not depart from the spirit and scope of this disclosure, and that various changes, substitutions, and alterations can be made to them without departing from the spirit and scope of this disclosure.

[0034] 100: Utilizing artificial intelligence for nitrogen dioxide gas detection and reporting systems 110: Image Collector 112: Monitor lens 114: Monitor host 120: Calculation and Analysis Server 122: Artificial Intelligence Image Analysis Module 124: Image Correction Module 126: Feature Extraction Module 128: Storage Module 130: Application terminal device A: Flue opening B: Nitrogen dioxide, a yellow gas C: Real-time smoke extraction video D: Identification Area S11~S14: Steps S21~S25: Steps

Claims

1. An artificial intelligence-based system for detecting and reporting nitrogen dioxide gas, for detecting yellow nitrogen dioxide gas (B) emitted from a flue (A), comprising: an image collector (110) having a monitor lens (112) and a monitor host (114) electrically connected to the monitor lens (112), wherein the monitor lens (112) is disposed on one side of the flue (A) and is used to capture images of the flue (A), and the monitor host (114) is used to acquire a real-time smoke emission image (C) generated by the monitor lens (112) capturing images of the flue (A); A computational analysis server (120) is communicatively connected to the image collector (110) and is used to receive the real-time smoke exhaust image (C) transmitted by the monitor host (114) of the image collector (110). The computational analysis server (120) has an artificial intelligence image analysis module (122), which is used to detect the real-time smoke exhaust image (C) according to an object detection model to generate analysis information and to track the analysis information according to a trajectory tracking model. The system tracks and generates a smoke identification result. The artificial intelligence image analysis module (122) further analyzes the smoke identification result according to an image analysis model to generate nitrogen dioxide gas concentration information. An application terminal device (130) is connected to the computing analysis server (120) and is used to receive the smoke identification result and the nitrogen dioxide gas concentration information transmitted by the computing analysis server (120). The application terminal device (130) is a personal computer, a laptop computer, a tablet computer, or a mobile device.

2. The AI-based nitrogen dioxide gas detection and reporting system described in claim 1 further includes: an image correction module (124) communicatively connected to the image collector (110) for image correction of the real-time smoke image (C) to obtain a plurality of preprocessing information.

3. The AI-based detection and reporting system for nitrogen dioxide gas as described in claim 2, wherein the computational analysis server (120) further comprises: a feature extraction module (126), communicatively connected to the image correction module (124), for receiving the plurality of preprocessing information and extracting a plurality of feature information from the plurality of preprocessing information; and a storage module (128), communicatively connected to the feature extraction module (126), for storing the plurality of feature information.

4. The artificial intelligence-based nitrogen dioxide gas detection and notification system as described in claim 2, wherein the object detection model is established based on the YOLO algorithm, and the artificial intelligence image analysis module (122) is further used to detect the plurality of preprocessed information after image correction of the real-time smoke exhaust image (C) based on the object detection model to generate the analysis information, and the trajectory tracking model used by the artificial intelligence image analysis module (122) of the computing analysis server (120) is a two-stage region-based convolutional neural network (R-CNN).

5. The artificial intelligence-based detection and notification system for nitrogen dioxide gas as described in claim 1, wherein the computational analysis server (120) is further configured to set an identification area (D) on the real-time smoke exhaust image (C), and the artificial intelligence image analysis module (122) is further configured to detect the identification area (D) according to the object detection model to generate the analysis information.

6. An operation method for detecting and reporting nitrogen dioxide gas using artificial intelligence, for detecting yellow nitrogen dioxide gas (B) emitted from a flue (A), comprising: capturing images of the flue (A) using a monitor lens (112) of an image collector (110), such that a monitor host (114) of the image collector (110) acquires a real-time smoke emission image (C) generated by capturing images of the flue (A), wherein the monitor lens (112) is electrically connected to the monitor host (114), and the monitor lens (112) is disposed on one side of the flue (A); receiving the real-time smoke emission image (C) transmitted by the monitor host (114) of the image collector (110) via a computing analysis server (120); An AI image analysis module (122) of the computing analysis server (120) detects the real-time smoke image (C) according to an object detection model to generate analysis information; the AI ​​image analysis module (122) of the computing analysis server (120) tracks the analysis information according to a trajectory tracking model to generate a smoke identification result; the AI ​​image analysis module (122) of the computing analysis server (120) analyzes the smoke identification result according to an image analysis model to generate nitrogen dioxide gas concentration information; and an application terminal device (130) connected to the computing analysis server (120) receives the smoke identification result and the nitrogen dioxide gas concentration information, wherein the application terminal device (130) is a personal computer, laptop, tablet computer or mobile device.

7. The method of using artificial intelligence to detect and report nitrogen dioxide gas as described in claim 6, wherein receiving the real-time smoke exhaust image (C) by the computing analysis server (120) further comprises: receiving the real-time smoke exhaust image (C) by an image correction module (124) of the computing analysis server (120), and performing image correction on the real-time smoke exhaust image (C) to obtain a plurality of preprocessing information.

8. The operation method of the artificial intelligence-based detection and reporting system for nitrogen dioxide gas as described in claim 7 further includes: receiving the plurality of preprocessed information by a feature extraction module (126) of the computational analysis server (120), and extracting a plurality of feature information from the plurality of preprocessed information; and storing the plurality of feature information by a storage module (128) of the computational analysis server (120).

Images