Distinguishing method and distinguishing system for illegal work in extreme weather

Abstract

The invention relates to a method and a system for judging violation work in extreme weather, and belongs to the technical field of work management and control. Comprising the following steps: periodically acquiring weather information in a specified range, when extreme weather occurs in a certain area in the specified range, calling audio and video information of a construction site in the area by a control platform, analyzing the audio and video information according to a video analysis model and / or a voiceprint analysis model, and sending the analyzed audio and video information to a server; and if the analysis finds that the construction operation under the extreme weather is early warned in time. According to the invention, by analyzing audio and video information of a construction site in a certain area where extreme weather such as storm wind, rainstorm and thunder occurs, construction work in the construction site in the area under the extreme weather is found in time, and early warning processing is carried out; according to the invention, the weather information and the audio and video information of the construction site are combined to judge whether construction work exists in the construction site and perform corresponding early warning processing, so that illegal work in extreme weather can be managed and controlled in time.

Description

Technical Field

[0001] The invention relates to a method and a system for distinguishing illegal operations in extreme weather, belonging to the technical field of operation control. Background Art

[0002] With the development of the economy and society, the scale of power grids has gradually expanded. The on-site power operation scenarios include infrastructure projects, overhaul projects, technical transformation projects, business expansion projects, etc., involving all aspects of generation, transmission, transformation and distribution. They are characterized by being numerous and extensive, and on-site safety management and control is difficult. At the same time, factors such as the uncertainty risks at the operation site and the substandard quality of the operators have put forward higher requirements for the intelligent management and control technology to ensure the safety of on-site operations.

[0003] At present, the safety risk identification technology for the construction process has gradually matured. It mainly uses image recognition and video analysis to accurately perceive personnel behavior and equipment status, timely discover dangerous factors in the operation process, and prevent accidents. Common technologies include target detection technology for static images and human skeleton recognition technology for time-series video analysis. They can identify high-frequency safety risks such as not wearing a safety helmet, not wearing insulating boots, unmanned escalators, falls, deep foundation pits, etc. However, this type of technology usually analyzes a single mode of video images and requires manual multi-faceted judgment to avoid illegal operations under multiple conditions.

[0004] There are usually certain requirements for weather conditions during power grid construction. In extreme weather, construction work needs to be prohibited to prevent safety accidents. Currently, there is little research on the identification of illegal operations in extreme weather based on intelligent sensing technology, and it mainly relies on manual identification. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for distinguishing illegal operations in extreme weather, so as to solve the problem that illegal operations in existing extreme weather cannot be timely controlled through manual distinction.

[0006] To achieve the above object, the solution of the present invention includes:

[0007] A method for distinguishing illegal operations in extreme weather of the present invention comprises the following steps:

[0008] Weather information within a specified range is obtained periodically. When extreme weather occurs in an area within the specified range, the control platform retrieves audio and video information from the construction site in the area and analyzes the audio and video information based on the video analysis model and / or voiceprint analysis model. If the analysis finds construction work under extreme weather conditions, timely warnings will be issued.

[0009] The present invention timely discovers the existence of construction work under extreme weather conditions at the construction site in a certain area where extreme weather conditions such as storms, rainstorms, and lightning occur by analyzing the audio and video information of the construction site in the area and performs early warning processing; the present invention combines the weather information and audio and video information of the construction site to determine whether there is construction work at the construction site and performs corresponding early warning processing, and timely controls illegal operations under extreme weather conditions. The designated range is determined by the location of the construction site that needs to be precisely controlled; the early warning processing includes early warning, and also includes linking the on-site camera to capture evidence.

[0010] Furthermore, the video analysis model includes any one or any combination of on-site construction personnel identification, construction vehicle identification and tower crane identification; on-site construction personnel identification, construction vehicle identification and tower crane identification all use target detection algorithms to obtain target IOUs, and target IOUs are used to count the number of on-site construction personnel to determine whether there are people stranded at the construction site; the offset of the relative position of the target IOU within a set period is used to determine whether the construction vehicles and tower cranes are moving.

[0011] The present invention obtains the target IOU by using the target detection algorithm on the audio and video information. The target IOU of the present invention is used to count the number of construction personnel on site to determine whether there are people stranded at the construction site. If the number of construction personnel on the construction site is greater than 0, it is considered that there are people stranded, that is, there are construction operations under extreme weather at the construction site; the offset of the relative position of the target IOU of the present invention within a set period is used to determine whether the construction vehicles and tower cranes are moving. The movement of construction vehicles and tower cranes belongs to construction operations under extreme weather. The target IOU is used to describe the degree of overlap between the corresponding position areas of the same type of target position information in the image to be detected.

[0012] Furthermore, the target detection algorithm is a target detection algorithm based on YOLOV5.

[0013] The target detection algorithm based on YOLOV5 adopted in the present invention makes the analysis conclusion of audio and video information more reliable.

[0014] Furthermore, when the period is set to 1 minute, the offset of the relative position of the target IOU within 1 minute is the difference between the target IOU of the first and last frames. If the difference between the target IOU of the first and last frames is less than the set value, it is considered that the construction vehicle and the tower crane are moving.

[0015] When the relative position offset of the target IOU of the present invention within 1 minute is less than the set value, it is considered that the construction vehicle and the tower crane are moving, that is, there is construction work under extreme weather conditions at the construction site.

[0016] Furthermore, the setting value ranges from 0.4 to 0.6.

[0017] Furthermore, the voiceprint analysis model is obtained in the following way: audio signals from the construction site are collected, Wiener filtering and Gaussian noise reduction are performed on the audio signals, frame division and windowing are performed according to the set time, and then time-frequency domain transformation and Mel cepstral coefficient feature extraction are performed. Finally, a binary classification of construction work and no construction work is performed based on the SVM binary classification algorithm model to determine whether there is construction work in extreme weather in a certain area.

[0018] The present invention collects audio signals from the construction site, performs Wiener filtering and Gaussian noise reduction on the audio signals, filters out wind, rain, bird calls, and lightning noise, and then performs frame division and windowing processing according to the set time. The window adopts a Hamming window, and then performs time-frequency domain transformation and Mel cepstral coefficient feature extraction. Finally, based on the SVM binary classification algorithm model, it is classified into construction work and no construction work to determine whether there is construction work in a certain area under extreme weather conditions.

[0019] Furthermore, the time is set to 0.02 seconds.

[0020] The present invention divides frames according to the set time, taking 0.02 seconds as one frame, and provides a reliable basis for time-frequency domain transformation and extraction of Mel-frequency cepstral coefficient features.

[0021] Furthermore, it also includes synchronizing weather information of extreme weather in a certain area to the control platform in real time, and the control platform prohibits the issuance of work tickets in a certain area when extreme weather occurs.

[0022] The present invention synchronizes weather information of extreme weather in a certain area to a control platform in real time, and prohibits the issuance of work tickets in a certain area during the time when extreme weather occurs through the control platform, thereby promptly blocking the occurrence of construction work under extreme weather from the source, and providing more accurate management and control of illegal operations under extreme weather.

[0023] Furthermore, it also includes the control platform automatically associating the information of work tickets that have been issued in a certain area, and issuing early warnings for work tickets that have not been completed when extreme weather occurs.

[0024] The control platform of the present invention can also automatically associate the information of work tickets that have been issued in a certain area, and issue early warnings for work tickets that have not been completed when extreme weather occurs. It can also issue early warnings for work tickets that have been issued and have not been completed when extreme weather occurs, and promptly stop construction work in extreme weather, providing more accurate control over illegal work in extreme weather.

[0025] The present invention provides a system for distinguishing illegal operations in extreme weather, including a processor, and the processor is used to execute instructions to implement the above-mentioned method for distinguishing illegal operations in extreme weather.

[0026] The present invention timely discovers the existence of construction work under extreme weather conditions at the construction site in a certain area where extreme weather conditions such as storms, rainstorms, and lightning occur by analyzing the audio and video information of the construction site in the area and performs early warning processing; the present invention combines the weather information and the audio and video information of the construction site to determine whether there is construction work at the construction site and performs corresponding early warning processing, thereby being able to timely control illegal operations under extreme weather conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flowchart of the present invention for distinguishing illegal operations in extreme weather;

[0028] Figure 2 It is a flowchart of the present invention for analyzing audio and video information based on a video analysis model and a voiceprint analysis model. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0030] An embodiment of a method for distinguishing illegal operations in extreme weather:

[0031] A method for distinguishing illegal operations in extreme weather includes the following steps:

[0032] Periodically obtain weather information within the specified range. When extreme weather occurs in a certain area within the specified range, the weather information of extreme weather in a certain area is synchronized to the control platform in real time. The control platform prohibits a certain area from issuing work tickets during the time of extreme weather. The control platform also automatically associates the work ticket information that has been issued in a certain area, and issues early warnings to the work tickets that have not been completed when extreme weather occurs. The control platform also selects the audio and video address in the area, retrieves the audio and video information of the construction site in the area by obtaining the audio and video stream, and analyzes the audio and video information according to the video analysis model and / or voiceprint analysis model. If the analysis finds that the construction work is under extreme weather, timely early warning will be issued. The specified range is determined by the location of the construction site that needs to be precisely controlled; the early warning process includes early warning, and also includes linking the on-site camera to capture evidence.

[0033] Among them, the video analysis model includes any one or any combination of on-site construction personnel identification, construction vehicle identification and tower crane identification; on-site construction personnel identification, construction vehicle identification and tower crane identification all use the target detection algorithm based on YOLOV5 to obtain the target IOU, and the target IOU is used to count the number of on-site construction personnel to determine whether there are people stranded at the construction site; the offset of the relative position of the target IOU within the set period is used to determine whether the construction vehicles and tower cranes are moving. The setting value range is 0.4 to 0.6. When the set period is 1 minute and the set value is 0.5, the offset of the relative position of the target IOU within 1 minute is the difference between the target IOU of the first and last frames. If the difference between the target IOU of the first and last frames is less than 0.5, it is considered that the construction vehicle and tower crane are moving.

[0034] Among them, the voiceprint analysis model is obtained in the following way: collect audio signals from the construction site, perform Wiener filtering and Gaussian noise reduction on the audio signals, filter out wind, rain, bird calls, and lightning noise, and then perform frame division and windowing processing according to the set time. The window adopts a Hamming window, and then performs time-frequency domain transformation and Mel cepstral coefficient feature extraction. Finally, based on the SVM binary classification algorithm model, a binary classification of construction work and no construction work is performed to determine whether there is construction work in extreme weather in a certain area. The present invention performs frame division according to the set time, taking 0.02 seconds as a frame, and provides a reliable basis for time-frequency domain transformation and Mel cepstral coefficient feature extraction.

[0035] like Figure 1 As shown, firstly, the weather conditions (weather information) within a certain range (specified range) are periodically obtained. When extreme weather such as storms, rainstorms, lightning, etc. occurs in a certain area (a certain area) within the certain range, the information is promptly synchronized to the e-infrastructure platform (the control platform adopts the e-infrastructure platform), prohibiting the area from opening work tickets during the time when extreme weather occurs, and automatically linking the information of the opened work tickets, issuing early warnings for the work tickets in the area, and at the same time retrieving the audio and video information of the construction site in the extreme weather area, performing audio and video analysis based on audio-visual fusion technology, and issuing timely early warnings for illegal operations in extreme weather, and linking the on-site cameras to capture evidence.

[0036] in Figure 1 The audio-visual fusion technology in includes video analysis model and voiceprint analysis model, such as Figure 2As shown in the figure, the details are as follows: First, construct a video analysis algorithm model (video analysis model) to achieve three aspects of identification: one is the identification of on-site construction personnel, counting the number of on-site construction personnel; the second is the identification of construction machinery and vehicles, judging whether the vehicle is moving; the third is the identification of tower cranes, judging whether there is mechanical rotation. Pedestrian recognition is based on the YOLOV5 target detection algorithm, mechanical vehicle movement recognition is based on YOLOV5 target detection and relative position offset, with a period of 1 minute. When the IOU of the first and last two frames of the target is less than 0.5, it is judged as mechanical vehicle movement. Tower crane rotation recognition is based on YOLOV5 target detection and relative position offset. It is also based on a period of 1 minute. When the IOU of the first and last two frames of the target is less than 0.5, it is judged as tower crane movement. When pedestrians, mechanical vehicles or tower cranes are moving on site, it is judged that there is illegal operation under extreme weather conditions, and an early warning is immediately issued and captured for evidence. Second, construct a voiceprint recognition algorithm model (voiceprint analysis model). Based on the audio signal collected on site, first perform Wiener filtering and Gaussian noise reduction to filter out wind, rain, bird calls, and lightning noise, then perform frame segmentation and windowing processing, generally taking 0.02s as a frame, using a Hamming window; then perform time-frequency domain transformation and Mel cepstral coefficient feature extraction; finally, based on the SVM binary classification algorithm model, implement binary classification of construction work and no construction work to determine whether there is illegal work on site. Generally, when there is construction work on site, the audio-visual channel can identify the result at the same time. Electric power construction takes safety as the primary principle. Logically, if the audio-visual recognizes that there is construction work, it is judged as illegal work in extreme weather.

[0037] The present invention firstly periodically obtains the weather conditions within a certain range. When extreme weather such as storms, heavy rains, lightning, etc. occurs in a certain area, the information is synchronized to the e-infrastructure platform in time, and the area is prohibited from issuing work tickets during the time when extreme weather occurs. The information of the work tickets that have been issued is automatically associated, and an early warning is issued for the work tickets in the area. At the same time, the audio and video information of the construction site in the extreme weather area is retrieved, and audio and video analysis is performed based on audio-visual fusion technology. When illegal operations are found in extreme weather, timely early warnings are issued and the on-site cameras are linked to capture evidence.

[0038] An embodiment of a system for distinguishing illegal operations in extreme weather:

[0039] A system for distinguishing illegal operations in extreme weather includes a processor, and the processor is used to execute instructions to implement the above-mentioned method for distinguishing illegal operations in extreme weather. The method for distinguishing illegal operations in extreme weather has been described in detail in an embodiment of a method for distinguishing illegal operations in extreme weather, and will not be repeated here.

Claims

1. A method for distinguishing illegal operations in extreme weather, characterized in that: The steps include: Weather information within a specified range is obtained periodically. When extreme weather occurs in an area within the specified range, the control platform retrieves audio and video information from the construction site in the area and analyzes the audio and video information based on the video analysis model and / or voiceprint analysis model. If the analysis finds construction work under extreme weather conditions, timely warnings will be issued.

2. The method for distinguishing illegal operations in extreme weather according to claim 1, characterized in that: The video analysis model includes any one or any combination of on-site construction personnel identification, construction vehicle identification and tower crane identification; the on-site construction personnel identification, construction vehicle identification and tower crane identification all use target detection algorithms to obtain target IOUs, and the target IOUs are used to count the number of on-site construction personnel to determine whether there are people stranded at the construction site; the relative position offset of the target IOU within a set period is used to determine whether the construction vehicles and tower cranes are moving.

3. The method for distinguishing illegal operations in extreme weather according to claim 2, characterized in that: The target detection algorithm is a target detection algorithm based on YOLOV5.

4. The method for distinguishing illegal operations in extreme weather according to claim 2 or 3, characterized in that: When the set period is 1 minute, the offset of the relative position of the target IOU within 1 minute is the difference between the target IOU of the first and last frames. If the difference between the target IOU of the first and last frames is less than the set value, it is considered that the construction vehicle and the tower crane are moving.

5. The method for distinguishing illegal operations in extreme weather according to claim 4, characterized in that: The setting value ranges from 0.4 to 0.

6.

6. The method for distinguishing illegal operations in extreme weather according to claim 1, characterized in that: The voiceprint analysis model is obtained in the following manner: audio signals from the construction site are collected, Wiener filtering and Gaussian noise reduction are performed on the audio signals, frame-by-frame windowing is performed according to a set time, time-frequency domain transformation and Mel cepstral coefficient feature extraction are performed, and finally, a binary classification of construction work and no construction work is performed based on the SVM binary classification algorithm model to determine whether there is construction work in extreme weather in the certain area.

7. The method for distinguishing illegal operations in extreme weather according to claim 6, characterized in that: The setting time is 0.02 seconds.

8. The method for distinguishing illegal operations in extreme weather according to claim 1, characterized in that: It also includes synchronizing weather information of extreme weather occurring in the certain area to a control platform in real time, and the control platform prohibits the issuance of work tickets in the certain area when extreme weather occurs.

9. The method for distinguishing illegal operations in extreme weather according to claim 1 or 8, characterized in that: It also includes the control platform automatically associating the work ticket information that has been issued in the certain area, and issuing early warnings for work tickets that have not been completed when extreme weather occurs.

10. A system for distinguishing illegal operations in extreme weather, characterized in that: It includes a processor, which is used to execute instructions to implement the method for distinguishing illegal operations in extreme weather as described in any one of claims 1 to 9.

Images