CN116797967A - Visual monitoring hidden trouble identification method and device for overhead transmission line - Google Patents

Visual monitoring hidden trouble identification method and device for overhead transmission line Download PDF

Info

Publication number
CN116797967A
CN116797967A CN202310548207.1A CN202310548207A CN116797967A CN 116797967 A CN116797967 A CN 116797967A CN 202310548207 A CN202310548207 A CN 202310548207A CN 116797967 A CN116797967 A CN 116797967A
Authority
CN
China
Prior art keywords
alarm
transmission line
smoke
judging
overhead transmission
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202310548207.1A
Other languages
Chinese (zh)
Inventor
刘焕云
蔡富东
吕昌峰
郭国信
帅民伟
丁健配
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Shandong Senter Electronic Co Ltd
Original Assignee
Shandong Senter Electronic Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shandong Senter Electronic Co Ltd filed Critical Shandong Senter Electronic Co Ltd
Priority to CN202310548207.1A priority Critical patent/CN116797967A/en
Publication of CN116797967A publication Critical patent/CN116797967A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G—PHYSICS
    • G08—SIGNALLING
    • G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
    • G08B21/00—Alarms responsive to a single specified undesired or abnormal condition and not otherwise provided for
    • G08B21/18—Status alarms
    • G08B21/182—Level alarms, e.g. alarms responsive to variables exceeding a threshold

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Computing Systems (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Health & Medical Sciences (AREA)
  • Databases & Information Systems (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Business, Economics & Management (AREA)
  • Emergency Management (AREA)
  • Closed-Circuit Television Systems (AREA)

Abstract

The invention discloses a method and a device for identifying hidden dangers of visual monitoring of an overhead transmission line, and belongs to the technical field of electric power. Firstly, detecting hidden danger of a periodically-captured image through a hidden danger identification model, judging as critical alarm when the hidden danger type is a wire foreign matter, and judging as general alarm, key alarm or critical alarm according to the movement trend of construction machinery or smoke mountain fire and the position relation between an alarm position and a protection area of an overhead transmission line when the hidden danger type is the construction machinery or the smoke mountain fire; for key alarms, carrying out timing intensive snapshot; and for critical alarms, collecting video streams of alarm positions, and analyzing the video streams. The invention reduces the workload of operation and maintenance personnel on alarm processing, avoids serious accidents such as tripping or casualties caused by untimely alarm processing, vehicle damage and the like, and can realize effective monitoring on hidden danger development and change processes while saving the electric quantity of equipment.

Description

Visual monitoring hidden trouble identification method and device for overhead transmission line
Technical Field
The invention relates to the technical field of electric power, in particular to a method and a device for identifying hidden dangers of visual monitoring of an overhead transmission line.
Background
At present, the hidden danger identification algorithm for overhead transmission line visualization is used for realizing hidden danger identification through a target detection scheme, a front-end device or a rear-end service sends an alarm to a platform after the hidden danger is identified, a monitoring person receives the alarm and decides subsequent arrangement according to the experience judgment of the alarm, the emergency alarm is solved through an audible and visual alarm or on-site inspection, the non-emergency alarm is kept to be closely tracked, no processing is performed, and the alarm which has no threat at all is not processed.
With the increasing number of visual monitoring devices, the acquisition vacuum period is shortened gradually, the alarm quantity on the platform is increased gradually, personnel monitoring is limited by the number of people, personnel fatigue and other conditions which possibly cause missed observation, so that emergency alarms are silenced in alarm messages, and serious accidents such as tripping, casualties, vehicle damage and the like are caused.
Disclosure of Invention
In order to solve the defects of the prior art, the invention provides the method and the device for identifying the hidden danger of the visual monitoring of the overhead transmission line, which reduce the workload of operation and maintenance personnel on alarm processing, avoid serious accidents such as tripping or casualties, vehicle damage and the like caused by untimely alarm processing, and can realize effective monitoring on the hidden danger development and change process while saving the electric quantity of equipment.
The technical scheme provided by the invention is as follows:
a method for identifying hidden dangers in visual monitoring of an overhead transmission line comprises the following steps:
s1: acquiring an image which is captured at regular time through a monitoring device, detecting hidden danger of the image through a trained hidden danger identification model, judging whether hidden danger exists or not, and judging the type of the hidden danger; wherein the hidden danger types comprise construction machinery, lead foreign matters and smoke forest fires;
s2: when the hidden danger type is judged to be construction machinery or smoke mountain fire, S3 is executed; when the hidden danger type is judged to be the foreign matter of the lead, judging the alarm level as critical alarm, and executing S4;
s3: judging the alarm level as a general alarm, a key alarm or a critical alarm according to the movement trend of the construction machine or the smoke forest fire and the position relation between the alarm position of the construction machine or the smoke forest fire and the set protection area of the overhead transmission line;
s4: for general alarms, the process is not performed, and the next patrol period is waited to be executed from S1; for important alarms, carrying out timing intensive snapshot through a monitoring device, and outputting important alarm information and images of the timing intensive snapshot; and for critical alarming, adjusting the position and focal length of the monitoring device, collecting video streams of the alarming position, analyzing the video streams, and outputting critical alarming information and analysis results of the video streams.
Further, the protection area of the overhead transmission line is set through the following process:
acquiring target images of a wire and a tower of an overhead transmission line, and detecting the tower and the wire of the target images through a tower detection model and a wire detection model respectively to obtain a tower position and a wire position;
determining a main tower according to the relative position information of the wire position and the tower position, determining the outermost wires at two sides of the main tower by using the wire position, and determining the area between the main tower and the outermost wires as an initial protection area;
and (3) expanding the initial protection area according to the voltage class of the overhead transmission line to obtain the protection area of the overhead transmission line.
Further, the initial guard area is expanded by the following process:
setting an expansion coefficient K and calculating an expansion distance R;
wherein, (r=x R -X L )*K/2，X R and XL The X coordinates of the right edge point and the X coordinates of the left edge point of the initial protection area are respectively;
for each contour point P of the initial protection zone i When the outline point P i When the points are convex, concave or flat, the outline point Q after expansion is calculated by the following formulas (1), (2) and (3) i ；
wherein ,
P i-1 and Pi+1 Respectively the contour points P i And a preceding contour point and a following contour point, θ i Is a line segment P i-1 P i And P i+ 1 P i Included angle e i Is P i-1 Pointing to P i Unit vector of e i+1 Is P i Pointing to P i+1 Is a unit vector of (a).
Further, the step S3 includes:
s31: calculating the intersection ratio IOU1 of a detection frame of the alarm position of the smoke mountain fire of the front image and the rear image, comparing the intersection ratio IOU1 with a set threshold T1, judging the alarm level as important alarm if the IOU1 is smaller than T1, otherwise executing S32;
calculating the intersection ratio IOU2 of a detection frame of the warning position of the construction machine of the front image and the rear image, comparing the intersection ratio IOU2 with a set threshold T2, judging the warning level as a general warning if the IOU2 is larger than T2, otherwise executing S32;
s32: judging whether an intersection exists between a detection frame of the warning position of the construction machine or the smoke mountain fire and a protection area of the overhead transmission line, if not, judging the warning grade as a general warning, and if so, executing S33;
s33: if the alarm position of the construction machine or the smoke forest fire is located outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is larger than the first distance, judging the alarm grade as a general alarm;
If the alarm position of the construction machine or the smoke forest fire is located outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is between the second distance and the first distance, judging the alarm level as a key alarm;
if the alarm position of the construction machine or the smoke forest fire is located in the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is smaller than the second distance, judging the alarm level as critical alarm;
wherein the second distance is less than the first distance.
Further, the video stream is analyzed by:
and detecting hidden danger of each frame of image of the video stream through a video detection model, continuously analyzing the detected hidden danger through a target tracking model, and calculating the motion state of the construction machine or the diffusion trend of smoke mountain fire through a trend judging method.
Further, the position of the monitoring device is adjusted by the following process:
calculating an X-direction adjustment scale angle_x and a Y-direction adjustment scale angle_y of the monitoring device, and adjusting the position of the monitoring device according to the angle_x and the angle_y;
Angle_x＝(bbox_center_x-image_center_x)*FOV_width/width
Angle_y＝(bbox_center_y-image_center_y)*FOV_height/height
wherein bbox_center_x and bbox_center_y are the X and Y coordinates of the center of the alert position, image_center_x and image_center_y are the X and Y coordinates of the center of the image, fov_width and fov_height are the width and height views of the monitor, respectively, and width and height are the width and height of the image, respectively;
The focal length of the monitoring device is adjusted through the following process:
calculating a New focal length New_focal of the monitoring device, and adjusting the focal length of the monitoring device according to the New focal length New_focal;
New_focal＝current_focal+bbox_width/width*scale
the current_focal is the current focal length of the monitor, bbox_width is the width of the monitor, and scale is the set scaling factor.
An overhead transmission line visual monitoring hidden trouble recognition device, the device comprising:
the hidden danger detection module is used for acquiring images which are captured at regular time through the monitoring device, detecting hidden danger of the images through the trained hidden danger identification model, judging whether hidden danger exists or not and judging the type of the hidden danger; wherein the hidden danger types comprise construction machinery, lead foreign matters and smoke forest fires;
the first level judging module is used for executing the second level judging module when the hidden danger type is judged to be construction machinery or smoke mountain fire; when the hidden danger type is judged to be the foreign matter of the lead, judging the alarm level as critical alarm, and executing an alarm processing module;
the second level judging module is used for judging the alarm level as a general alarm, a key alarm or a critical alarm according to the movement trend of the construction machinery or the smoke mountain fire and the position relation between the alarm position of the construction machinery or the smoke mountain fire and the set protection area of the overhead transmission line;
The alarm processing module is used for processing general alarms, and waiting for the execution of the next inspection period from the hidden danger detection module; for important alarms, carrying out timing intensive snapshot through a monitoring device, and outputting important alarm information and images of the timing intensive snapshot; and for critical alarming, adjusting the position and focal length of the monitoring device, collecting video streams of the alarming position, analyzing the video streams, and outputting critical alarming information and analysis results of the video streams.
Further, the protection area of the overhead transmission line is set through the following process:
acquiring target images of a wire and a tower of an overhead transmission line, and detecting the tower and the wire of the target images through a tower detection model and a wire detection model respectively to obtain a tower position and a wire position;
determining a main tower according to the relative position information of the wire position and the tower position, determining the outermost wires at two sides of the main tower by using the wire position, and determining the area between the main tower and the outermost wires as an initial protection area;
and (3) expanding the initial protection area according to the voltage class of the overhead transmission line to obtain the protection area of the overhead transmission line.
Further, the initial guard area is expanded by the following process:
setting an expansion coefficient K and calculating an expansion distance R;
wherein, (r=x R -X L )*K/2，X R and XL The X coordinates of the right edge point and the X coordinates of the left edge point of the initial protection area are respectively;
for each contour point P of the initial protection zone i When the outline point P i When the points are convex, concave or flat, the outline point Q after expansion is calculated by the following formulas (1), (2) and (3) i ；
wherein ,
P i-1 and Pi+1 Respectively the contour points P i And a preceding contour point and a following contour point, θ i Is a line segment P i-1 P i And P i+ 1 P i Included angle e i Is P i-1 Pointing to P i Unit vector of e i+1 Is P i Pointing to P i+1 Is a unit vector of (a).
Further, the second level judgment module includes:
the motion trend judging unit is used for calculating the intersection ratio IOU1 of the detection frames of the smoke mountain fire alarm positions of the front image and the rear image, comparing the intersection ratio IOU1 with a set threshold T1, judging the alarm level as a key alarm if the IOU1 is smaller than T1, and executing the qualitative analysis unit if the alarm level is not smaller than T1;
calculating the intersection ratio IOU2 of a detection frame of the warning position of the construction machine of the front image and the rear image, comparing the intersection ratio IOU2 with a set threshold T2, judging the warning grade as a general warning if the IOU2 is larger than the threshold T2, and executing a qualitative analysis unit if the IOU2 is not larger than the threshold T2;
The qualitative analysis unit is used for judging whether an intersection exists between a detection frame of the warning position of construction machinery or smoke mountain fire and a protection area of the overhead transmission line, if not, judging the warning grade as a general warning, and if so, executing the quantitative analysis unit;
the quantitative analysis unit is used for judging the alarm grade as a general alarm if the alarm position of the construction machine or the smoke mountain fire is located outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke mountain fire and the wire position is larger than a first distance;
if the alarm position of the construction machine or the smoke forest fire is located outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is between the second distance and the first distance, judging the alarm level as a key alarm;
if the alarm position of the construction machine or the smoke forest fire is located in the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is smaller than the second distance, judging the alarm level as critical alarm;
wherein the second distance is less than the first distance.
Further, the video stream is analyzed by:
And detecting hidden danger of each frame of image of the video stream through a video detection model, continuously analyzing the detected hidden danger through a target tracking model, and calculating the motion state of the construction machine or the diffusion trend of smoke mountain fire through a trend judging method.
Further, the position of the monitoring device is adjusted by the following process:
calculating an X-direction adjustment scale angle_x and a Y-direction adjustment scale angle_y of the monitoring device, and adjusting the position of the monitoring device according to the angle_x and the angle_y;
Angle_x＝(bbox_center_x-image_center_x)*FOV_width/width
Angle_y＝(bbox_center_y-image_center_y)*FOV_height/height
wherein bbox_center_x and bbox_center_y are the X and Y coordinates of the center of the alert position, image_center_x and image_center_y are the X and Y coordinates of the center of the image, fov_width and fov_height are the width and height views of the monitor, respectively, and width and height are the width and height of the image, respectively;
the focal length of the monitoring device is adjusted through the following process:
calculating a New focal length New_focal of the monitoring device, and adjusting the focal length of the monitoring device according to the New focal length New_focal;
New_focal＝current_focal+bbox_width/width*scale
the current_focal is the current focal length of the monitor, bbox_width is the width of the monitor, and scale is the set scaling factor.
A computer readable storage medium for overhead transmission line visual monitoring hidden trouble identification, comprising a memory for storing processor executable instructions which when executed by the processor implement steps comprising the overhead transmission line visual monitoring hidden trouble identification method.
The equipment for identifying the visual monitoring hidden trouble of the overhead transmission line comprises at least one processor and a memory for storing computer executable instructions, wherein the steps of the method for identifying the visual monitoring hidden trouble of the overhead transmission line are realized when the processor executes the instructions.
The invention has the following beneficial effects:
1. the invention divides the threat level of hidden danger into three levels, in particular to critical alarm, important alarm and general alarm. The critical alarms refer to alarms which operation and maintenance personnel need to intervene, the important alarms are alarms which are monitored in an important way, intervention preparation is carried out for the subsequent development of the critical alarms, and the general alarms are alarms which do not need attention and have no threat to wires and towers. The invention can analyze pertinently according to the type of hidden danger, fully excavates hidden danger characteristics of the service scene of the power transmission line to define effective alarms, critical alarms and important alarms are effective alarms which need to be concerned, and general alarms are invalid alarms which do not need to be concerned. The operation and maintenance personnel only need to pay attention to the important alarms and preferentially process the critical alarms, and do not need to pay attention to the general alarms. The workload of operation and maintenance personnel on alarm processing is reduced, the condition of missed observation of the alarm is avoided, and serious accidents such as tripping or casualties, vehicle damage and the like caused by the alarm are avoided.
2. Because of the power consumption requirement, the video analysis can not be started for a long time, and meanwhile, the identification precision of the video analysis model is generally much lower than that of the image analysis model, so the video analysis method is divided into two processes of image analysis and video analysis. The regular snapshot image of the monitoring device is normally analyzed, the capturing is intensively performed when the key alarm is judged, the video is collected and analyzed when the critical alarm is performed, and the effective monitoring on the hidden danger development change process can be realized while the electric quantity of equipment is saved by effective processing means under the key alarm and the critical alarm.
Drawings
FIG. 1 is a flow chart of a method for identifying hidden danger in visual monitoring of an overhead transmission line;
fig. 2 is a schematic diagram of an initial protection zone of an overhead transmission line;
FIG. 3 is a schematic illustration of initial guard area expansion;
FIG. 4 is a schematic illustration of a guard zone after initial guard zone expansion;
FIG. 5 is a schematic diagram of an IOU;
fig. 6 is a schematic diagram of the overhead transmission line visual monitoring hidden trouble recognition device of the present invention.
Detailed Description
In order to make the technical problems, technical solutions and advantages to be solved more clear, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings and specific embodiments. It will be apparent that the described embodiments are only some, but not all, embodiments of the invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be made by a person skilled in the art without making any inventive effort, are intended to be within the scope of the present invention.
The embodiment of the invention provides a method for identifying hidden dangers in visual monitoring of an overhead transmission line, which comprises the following steps of:
s1: acquiring images which are captured at regular time through a monitoring device, detecting hidden danger of the images through a trained hidden danger identification model, judging whether hidden danger exists or not, and judging the type of the hidden danger; wherein, hidden danger type includes construction machinery, wire foreign matter and smog mountain fire.
The analysis requirement of the hidden danger of the service scene of the power transmission line is that scene images are acquired through monitoring devices such as cameras and the like arranged on the tower of the power transmission line, an image is shot at regular intervals of 10-30 minutes, and the interval time is adjustable. For a 1.3 ten thousand-rack empty transmission line installed in a certain city, 1.3 ten thousand scenes are about 20%, an image is shot every 10 minutes according to the fact that hidden danger exists, 6 images are required to be acquired in one hour, and the alarm probability of 1.3 ten thousand scenes is 1.3×6×0.2=1.56 ten thousand. Every 10 minutes, 1.3 ten thousand 0.2=2600 hidden trouble images are required to be checked, a general conventional operation and maintenance team is responsible for about less than 10 people with 1.3 ten thousand devices, each person needs to be responsible for 260 scenes, the threat level is required to be confirmed besides the alarm, long-time monitoring is required, and a spot is visible on the fatigue level.
For certain scenes, the hidden trouble is motionless or moves in a region far away from the lead, no threat is formed to the tower, and the hidden trouble is not necessary to be reported to operation and maintenance personnel to follow up; however, for some hidden troubles, such as foreign matters on wires, hanging on a pole tower is dangerous, and the hidden troubles must be treated as soon as possible; for fireworks, under the condition of smaller and controllable fire, only attention is needed, and follow-up treatment is not needed; for cranes, it is not necessary to push to the service personnel even if the distance from the wire is large. Therefore, it is necessary to classify the threat level of the hidden trouble, and judge whether to push to the operation and maintenance personnel according to the level of the hidden trouble, so as to reduce the workload of the operation and maintenance personnel.
Because the hidden danger level is related to the hidden danger type, and the hidden danger level judging modes of different hidden danger types are different, the hidden danger position judging method and device firstly detect hidden danger of the images captured at fixed time, judge whether hidden danger exists or not and judge the hidden danger type, and meanwhile position the hidden danger to obtain the hidden danger position. The location of the hidden trouble can be generally represented by a detection box.
Conventional hidden danger recognition models are various, such as hidden danger recognition models based on yolov5, based on Faster RCNN and based on CascadeRCNN, and the hidden danger recognition models need to be trained before use.
Taking a Yolov5 target detection model as an example, inputting marked hidden danger data (comprising three major categories of construction machinery, wire foreign matters and smoke mountain fires, wherein the construction machinery comprises a crane, a cement pump truck, a pile driver, a dump truck and the like) into a Yolov5 network, performing Mosaic enhancement, anchor frame adaptive calculation and image scaling on training data, inputting the training data into a backbone network to extract image features, finally inputting the training data into an FPN and PAN structure to fuse multi-scale features, finally inputting the training data into a loss function to calculate loss, and back-propagating errors, thereby realizing the parameter training process of the model.
And detecting hidden danger of the input image after training, and if no hidden danger is identified, continuing to wait for the next image acquisition at regular time. If the hidden danger is detected, a detection frame at the hidden danger position is provided, the target type is judged, and the detection frame is judged to be construction machinery, lead foreign matters and smoke mountain fire.
S2: s3, executing when the hidden danger type is judged to be construction machinery or smoke mountain fire; when the hidden trouble type is judged as the foreign matter of the wire, the alarm level is judged as the critical alarm, and S4 is executed.
Different types of alarms are classified in different ways because of different degrees of threat. If the wire is foreign, the wire is threatened to run safely at any time and needs to be processed urgently, so that critical alarms are directly judged, and operation and maintenance personnel are reminded of needing to be processed urgently until the alarm is eliminated; in the case of pyrotechnical and construction machine types, further analysis is required.
S3: and judging the alarm grade as a general alarm, a key alarm or a critical alarm according to the movement trend of the construction machine or the smoke mountain fire and the position relation between the alarm position of the construction machine or the smoke mountain fire and the set protection area of the overhead transmission line.
The movement trend and the relative position of construction machinery or smoke forest fire can cause different threats to the overhead transmission line, so that the alarm level needs to be judged by combining the movement trend and the relative position. The types of smoke and construction machine are different, but the judgment logic is similar.
S4: for general alarms, the general alarms are not processed, and the general alarms do not need to be reported to operation and maintenance personnel to follow up, and the next inspection period is waited to be executed from S1.
And for the key alarm, the monitoring device is used for carrying out timing intensive snapshot, and key alarm information and images of the timing intensive snapshot are output.
After the key alarm is started, the key alarm is fed back to the monitoring device for intensive snapshot, and further development and change of hidden danger are closely observed. The monitoring rule of image analysis is generally that the image is shot once in 10-30 minutes, and the acquisition interval of intensive monitoring can be shortened to 1-5 minutes. This interval is only an example and may be actually set according to circumstances. And after the intensive snapshot, reporting the grade of the key alarm and the image of the intensive snapshot to operation and maintenance personnel for follow-up.
For critical alarms, the position and focal length of the monitoring device are adjusted, video streams of the alarm position are collected, the video streams are analyzed, and critical alarm information and analysis results of the video streams are output.
And after the critical alarm is started, adjusting the camera holder and the focal length of the camera, and starting a real-time video analysis or silent video analysis scheme. The purpose of the adjusting scheme of the camera cradle head and the camera focal length is to keep the hidden danger target always positioned in the middle of the picture and large enough. The real-time video analysis scheme and the silent video analysis scheme are used for further detecting hidden danger of the acquired video stream data and performing tighter analysis on the hidden danger. The hidden danger identification method of the video analysis is similar to the hidden danger analysis of the picture, and after the hidden danger identification method is analyzed, the dangerous alarm information and the analysis result of the video stream are reported to operation and maintenance personnel for intervention. The analysis result of the video stream may be in different forms of real-time video, short video of sample composition, dense image sequence, etc.
The invention has the following beneficial effects:
1. the invention divides the threat level of hidden danger into three levels, in particular to critical alarm, important alarm and general alarm. The critical alarms refer to alarms which operation and maintenance personnel need to intervene, the important alarms are alarms which are monitored in an important way, intervention preparation is carried out for the subsequent development of the critical alarms, and the general alarms are alarms which do not need attention and have no threat to wires and towers. The invention can analyze pertinently according to the type of hidden danger, fully excavates hidden danger characteristics of the service scene of the power transmission line to define effective alarms, critical alarms and important alarms are effective alarms which need to be concerned, and general alarms are invalid alarms which do not need to be concerned. The operation and maintenance personnel only need to pay attention to the important alarms and preferentially process the critical alarms, and do not need to pay attention to the general alarms. The workload of operation and maintenance personnel on alarm processing is reduced, the condition of missed observation of the alarm is avoided, and serious accidents such as tripping or casualties, vehicle damage and the like caused by the alarm are avoided.
2. Because of the power consumption requirement, the video analysis can not be started for a long time, and meanwhile, the identification precision of the video analysis model is generally much lower than that of the image analysis model, so the video analysis method is divided into two processes of image analysis and video analysis. The regular snapshot image of the monitoring device is normally analyzed, the capturing is intensively performed when the key alarm is judged, the video is collected and analyzed when the critical alarm is performed, and the effective monitoring on the hidden danger development change process can be realized while the electric quantity of equipment is saved by effective processing means under the key alarm and the critical alarm.
The invention defines the protection area of the overhead transmission line, which can be defined empirically and can be set according to the image recognition method. By way of example, the protection zone of an overhead transmission line may be set by the following procedure:
1. and acquiring target images of the wires and the towers of the overhead transmission line, and detecting the towers and the wires of the target images through the tower detection model and the wire detection model respectively to obtain the positions of the towers and the wires.
The tower detection can adopt a method based on deep learning, the tower detection model uses MobileNet as a model backbone network, and the loss function is GFLoss (GeneralFocalLoss), specifically:
GFLoss＝QFL+γDFL
Wherein, the super parameter γ=0.25, qfl and DFL are as follows:
QFL(p)＝-|y-σ| β ((1-y)log(1-σ)+ylog(σ))
DFL(S i ,S i+1 )＝-((y i+1 -y)log(S i )+(y-y i )log(S i+1 ))
in QFL, y is 0 or 1, the values of the predicted frame and the real frame IoU are represented, sigma is the class predicted value of the classified branch after the classified branch is subjected to a sigmoid activation function, the range is 0-1, and beta is the super parameter set to 2. In the DFL, y is a prediction boundary value, y i In order to obtain the real frame value after the average division and the rounding, the value of the super parameter i is an integer of 1 to 16.
The wire detection part also adopts a deep learning method, the model part is consistent with the tower detection model, and the difference is that the loss function in the wire detection training process is different, and the wire detection loss function is CELoss.
2. And determining the main tower according to the relative position information of the wire positions and the tower positions, determining the outermost wires at the two sides of the main tower by using the wire positions, and determining the area between the main tower and the outermost wires as an initial protection area.
The number of towers in the image is multiple, and towers corresponding to the wires, namely main towers, need to be determined. The main shaft tower may be determined according to a positional relationship between the wire and the shaft tower as shown in the middle shaft tower of fig. 2, and in general, the wire is directly connected to the main shaft tower and is positioned close thereto, so that the main shaft tower may be determined by the position.
The outermost wire can be determined according to the slope of the wire, and because of the imaging characteristics of the camera, the wire is imaged in a central projection manner in the image, the slope of the inner wire is different from that of the outer wire, and the slope of the outer wire is smaller than that of the inner wire, as shown in fig. 2. Therefore, the outermost wires at both sides of the main tower can be determined according to the slopes of the wires.
Finally, an initial protection area can be determined according to the main pole tower and the outermost wires, and the initial protection area can be formed according to mapping between the wires at the two sides of the outermost sides and the bottom of the main pole tower, as shown in fig. 2.
3. And (5) expanding the initial protection area according to the voltage class of the overhead transmission line to obtain the protection area of the overhead transmission line.
Because the voltage grades of the overhead transmission lines are different, the safety areas are different, and the initial protection area can be expanded according to the voltage grades of the overhead transmission lines to obtain the final protection area.
The external expansion method can be manually adjusted through experience, and can also be automatically expanded according to the initial protection area.
By way of example, the initial guard zone may be flared by the following procedure:
1. and setting an expansion coefficient K and calculating an expansion distance R.
Wherein, (r=x R -X L )*K/2，X R and XL The X-coordinate of the right edge point and the X-coordinate of the left edge point of the initial guard zone, respectively.
2. For each contour point P of the initial protection zone i When the outline point P i When the points are convex, concave or flat, the outline point Q after expansion is calculated by the following formulas (1), (2) and (3) i 。
wherein ,
P i-1 and Pi+1 Respectively the contour points P i And a preceding contour point and a following contour point, θ i Is a line segment P i-1 P i And P i+ 1 P i Included angle e i Is P i-1 Pointing to P i Unit vector of e i+1 Is P i Pointing to P i+1 Is a unit vector of (a).
The convex points, concave points or flat points can be obtained by deriving from the vicinity of the contour points, and the expansion of the convex points, concave points or flat points is shown in (a), (b) and (c) of fig. 3. An example of the guard area after the expansion is shown in fig. 4, where the expansion coefficient k=1.5.
As one such aspect of the embodiment of the present invention, the aforementioned S3 includes:
s31: calculating the intersection ratio IOU1 of the detection frames of the alarm positions of the smoke and the mountain fire of the front image and the rear image, comparing the intersection ratio IOU1 with a set threshold T1, judging the alarm level as an important alarm if the IOU1 is smaller than T1, and executing S32 otherwise.
The overlap ratio IOU is used to represent the overlapping degree of two detection frames, and its schematic diagram is shown in fig. 5, and its formula is as follows:
the IOU1 is used for judging the fire trend of smoke, and the detection frame of the alarm position of the smoke mountain fire comprises information of flames and smoke areas, so that whether the fire is enlarged or not can be judged through the diffusion of the smoke. The intersection ratio IOU1 of the two detection results is deduced, if the intersection ratio IOU1 is smaller than T1, the important alarm is carried out to remind operation and maintenance personnel of continuously paying attention. If the intersection ratio < T1 is considered, the qualitative and quantitative analysis of S32 is performed without enlarging the fire.
Calculating the intersection ratio IOU2 of the detection frames of the alarm positions of the construction machinery of the front image and the rear image, comparing the intersection ratio IOU2 with a set threshold T2, judging the alarm level as a general alarm if the IOU2 is larger than the threshold T2, and executing S32 otherwise.
And (3) carrying out motion trend judgment on the construction machine through the intersection ratio IOU2, carrying out qualitative or quantitative analysis of S32 if the IOU2 is smaller than T2 and carrying out general warning if the IOU2 is larger than T2 and the construction machine has almost no motion when the construction machine alarms similar to the smoke mountain fire.
The values of the thresholds T1 and T2 are adjustable, and may be, for example, 0.8.
S32: and judging whether an intersection exists between a detection frame of the warning position of the construction machine or the smoke mountain fire and a protection area of the overhead transmission line, if not, judging the warning grade as a general warning, and if so, executing S33.
The step is used for firstly carrying out qualitative analysis, judging whether the alarm position of the construction machine or the smoke mountain fire is in the protection area or not, and the judging method can use the position relation between the smoke detection frame and the position frame of the protection area to judge, specifically, judging whether the intersection exists between the detection frame and the protection area, if the intersection exists, the construction machine or the smoke mountain fire is considered to be at the edge or inside the protection area, and carrying out quantitative analysis of S33. If no intersection exists, the hidden danger is considered to be far away from the protection area, and a general alarm is fed back.
S33: and if the alarm position of the construction machine or the smoke forest fire is positioned outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is larger than the first distance, judging the alarm grade as a general alarm. Wherein the second distance is less than the first distance.
For example, if smoke or machinery occurs at the edge of the guard and the nearest distance from the wire exceeds 50 meters, a general alarm is given.
And if the alarm position of the construction machine or the smoke forest fire is located outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is between the second distance and the first distance, judging the alarm level as a key alarm.
For example, an important alarm is given if smoke or construction machinery occurs at the edge of the guard area or the clearance distance is between 30-50 meters.
And if the alarm position of the construction machine or the smoke forest fire is positioned in the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is smaller than the second distance, judging the alarm level as critical alarm.
For example, if smoke or construction machinery is inside the protected area or the clearance distance from the wire is less than 30 meters, a critical alarm is given. The above range can be set according to practical requirements, and is not limited to two numbers of 30 and 50.
The invention designs an effective alarm analysis method, utilizes the multidimensional characteristic indexes of hidden danger, including historical alarm indexes, current alarm indexes, target motion state analysis, qualitative and quantitative estimation and the like, divides alarm grades, preferentially sends critical alarms to monitoring staff, tracks the critical alarms in real time, records the process of the critical alarms while realizing timely early warning of the hidden danger, avoids tripping accidents, realizes the tracing of the hidden danger process, and protects the safe and stable operation of a power transmission line.
As another improvement of the embodiment of the present invention, the video stream may be analyzed by the following procedure:
each frame of image in the video stream is obtained, hidden danger of each frame of image in the video stream is detected through a video detection model, then a detected hidden danger target is subjected to continuous analysis through a target tracking model, and then the motion state of the construction machine or the spreading trend of smoke and mountain fire is calculated through a trend judging method.
The basic methods on which the present invention depends, such as a target detection method, a target tracking method, a smoke diffusion method, and a construction machine motion analysis method, are not particularly limited. For example, the target tracking algorithm may be a common single target, multi-target algorithm such as KCF, sort, deepSort. The analysis of the motion state of the construction machine may be performed by using a method based on an optical flow, a method based on movement of a target frame, or the like. The real-time video analysis method, that is, analyzing not less than 15 frames of images per second, and the silent video analysis method may analyze less than 10 images for 1s, for example, 1s for 1.
Further, the position of the monitoring device can be adjusted by the following process:
and calculating an X-direction adjustment scale angle_x and a Y-direction adjustment scale angle_y of the monitoring device, and adjusting the position of the monitoring device according to the angle_x and the angle_y.
Angle_x＝(bbox_center_x-image_center_x)*FOV_width/width
Angle_y＝(bbox_center_y-image_center_y)*FOV_height/height
Wherein bbox_center_x and bbox_center_y are the X and Y coordinates of the center of the alarm position, image_center_x and image_center_y are the X and Y coordinates of the center of the image, fov_width and fov_height are the width and height views of the monitor, respectively, and can be lookup tables, with width and height being the width and height of the image, respectively. For cameras with non-uniform distribution of viewing angles and pixels, the offset angles corresponding to the pixels can be calculated more accurately by using a lookup table.
Accordingly, the focal length of the monitoring device can be adjusted by the following process:
and calculating a New focal length New_focal of the monitoring device, and adjusting the focal length of the monitoring device according to the New focal length New_focal.
New_focal＝current_focal+bbox_width/width*scale
The current_focal is the current focal length of the monitor, bbox_width is the width of the monitor, and scale is the set scaling factor. scale is related to the position of the bottom of the object, and the closer the bottom of the object is to the upper part of the image, the larger the proportion of the object to the camera is, and different camera parameters are different.
The embodiment of the invention also provides a device for identifying hidden dangers in the visualized monitoring of the overhead transmission line, as shown in fig. 6, which comprises the following steps:
the hidden danger detection module 1 is used for acquiring images which are captured at regular time through the monitoring device, detecting hidden danger of the images through the trained hidden danger identification model, judging whether hidden danger exists or not and judging the type of the hidden danger; wherein, hidden danger type includes construction machinery, wire foreign matter and smog mountain fire.
A first level judgment module 2 for executing the second level judgment module 3 when the hidden danger type is judged as construction machinery or smoke mountain fire; when the hidden trouble type is judged as the wire foreign matter, the alarm level is judged as a critical alarm, and the alarm processing module 4 is executed.
The second level judging module 3 is configured to judge the alarm level as a general alarm, a key alarm or a critical alarm according to the movement trend of the construction machine or the smoke mountain fire and the positional relationship between the alarm position of the construction machine or the smoke mountain fire and the set protection area of the overhead transmission line.
The alarm processing module 4 is used for carrying out no processing on the general alarm and waiting for the execution of the next inspection period from the hidden danger detection module 1; for important alarms, carrying out timing intensive snapshot through a monitoring device, and outputting important alarm information and images of the timing intensive snapshot; for critical alarms, the position and focal length of the monitoring device are adjusted, video streams of the alarm position are collected, the video streams are analyzed, and critical alarm information and analysis results of the video streams are output.
The invention defines the protection area of the overhead transmission line, which can be defined empirically and can be set according to the image recognition method. By way of example, the protection zone of an overhead transmission line may be set by the following procedure:
and acquiring target images of the wires and the towers of the overhead transmission line, and detecting the towers and the wires of the target images through the tower detection model and the wire detection model respectively to obtain the positions of the towers and the wires.
And determining the main tower according to the relative position information of the wire positions and the tower positions, determining the outermost wires at the two sides of the main tower by using the wire positions, and determining the area between the main tower and the outermost wires as an initial protection area.
And (5) expanding the initial protection area according to the voltage class of the overhead transmission line to obtain the protection area of the overhead transmission line.
Further, the initial guard area may be flared by the following procedure:
and setting an expansion coefficient K and calculating an expansion distance R.
Wherein, (r=x R -X L )*K/2，X R and XL X coordinates of the right edge point and X coordinates of the left edge point of the initial guard zone, respectivelyCoordinates.
For each contour point P of the initial protection zone i When the outline point P i When the points are convex, concave or flat, the outline point Q after expansion is calculated by the following formulas (1), (2) and (3) i 。
wherein ,
P i-1 and Pi+1 Respectively the contour points P i And a preceding contour point and a following contour point, θ i Is a line segment P i-1 P i And P i+ 1 P i Included angle e i Is P i-1 Pointing to P i Unit vector of e i+1 Is P i Pointing to P i+1 Is a unit vector of (a).
As an improvement of the embodiment of the present invention, the aforementioned second level judgment module includes:
and the motion trend judging unit is used for calculating the intersection ratio IOU1 of the detection frames of the smoke and mountain fire alarm positions of the front image and the rear image, comparing the intersection ratio IOU1 with a set threshold T1, judging the alarm level as a key alarm if the IOU1 is smaller than T1, and executing the qualitative analysis unit if the alarm level is not smaller than T1.
Calculating the intersection ratio IOU2 of the detection frames of the alarm positions of the construction machinery of the front image and the rear image, comparing the intersection ratio IOU2 with a set threshold T2, judging the alarm level as a general alarm if the IOU2 is larger than the threshold T2, and executing a qualitative analysis unit if the alarm level is not larger than the threshold T2.
And the qualitative analysis unit is used for judging whether the intersection exists between the detection frame of the alarm position of the construction machine or the smoke mountain fire and the protection area of the overhead transmission line, if not, judging the alarm grade as a general alarm, and if so, executing the quantitative analysis unit.
And the quantitative analysis unit is used for judging the alarm grade as a general alarm if the alarm position of the construction machine or the smoke mountain fire is positioned outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke mountain fire and the wire position is larger than the first distance.
And if the alarm position of the construction machine or the smoke forest fire is located outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is between the second distance and the first distance, judging the alarm level as a key alarm.
And if the alarm position of the construction machine or the smoke forest fire is positioned in the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is smaller than the second distance, judging the alarm level as critical alarm.
Wherein the second distance is less than the first distance.
As another improvement of the embodiment of the present invention, the video stream may be analyzed by the following procedure:
the hidden danger of each frame of image of the video stream is detected through a video detection model, the detected hidden danger is continuously analyzed through a target tracking model, and the motion state of the construction machine or the spreading trend of smoke and mountain fire is calculated through a trend judging method.
Further, the position of the monitoring device can be adjusted by the following process:
and calculating an X-direction adjustment scale angle_x and a Y-direction adjustment scale angle_y of the monitoring device, and adjusting the position of the monitoring device according to the angle_x and the angle_y.
Angle_x＝(bbox_center_x-image_center_x)*FOV_width/width
Angle_y＝(bbox_center_y-image_center_y)*FOV_height/height
Wherein bbox_center_x and bbox_center_y are the X and Y coordinates of the center of the alert position, respectively, image_center_x and image_center_y are the X and Y coordinates of the center of the image, fov_width and fov_height are the width and height views of the monitor, respectively, and width and height are the width and height of the image, respectively.
Accordingly, the focal length of the monitoring device can be adjusted by the following process:
and calculating a New focal length New_focal of the monitoring device, and adjusting the focal length of the monitoring device according to the New focal length New_focal.
New_focal＝current_focal+bbox_width/width*scale
The current_focal is the current focal length of the monitor, bbox_width is the width of the monitor, and scale is the set scaling factor.
The device provided in the foregoing embodiments has a one-to-one correspondence between its implementation principle and the technical effects that are produced and the embodiments of the foregoing methods, and for a brief description, reference may be made to the corresponding matters in the embodiments of the foregoing methods where no mention is made in the examples of the device. It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of modules and units described in the apparatus may refer to corresponding procedures in the foregoing method embodiments, and are not described herein again.
The method according to the above embodiment of the present invention may implement service logic through a computer program and record the service logic on a storage medium, where the storage medium may be read and executed by a computer, so as to implement the effects of the solution described in the method embodiment of the present specification. Therefore, the embodiment of the invention also provides a computer readable storage medium for identifying the visual monitoring hidden trouble of the overhead transmission line, which comprises a memory for storing instructions executable by a processor, wherein the instructions realize the steps of the method for identifying the visual monitoring hidden trouble of the overhead transmission line.
The storage medium may include physical means for storing information, typically by digitizing the information before storing it in an electronic, magnetic, or optical medium. The storage medium may include: means for storing information using electrical energy such as various memories, e.g., RAM, ROM, etc.; devices for storing information using magnetic energy such as hard disk, floppy disk, magnetic tape, magnetic core memory, bubble memory, and USB flash disk; devices for optically storing information, such as CDs or DVDs. Of course, there are other ways of readable storage medium, such as quantum memory, graphene memory, etc.
The above description of the storage medium according to the method embodiment may further include other implementations, and the implementation principle and the generated technical effects of the embodiment are the same as those of the foregoing method embodiment, and specific reference may be made to the description of the related method embodiment, which is not repeated herein.
The embodiment of the invention also provides equipment for identifying the hidden trouble of the visual monitoring of the overhead transmission line, which can be a single computer or can comprise an actual operating device and the like using one or more of the methods or one or more of the embodiment devices of the specification. The equipment for identifying the visual monitoring hidden danger of the overhead transmission line can comprise at least one processor and a memory for storing computer executable instructions, wherein the steps of the method for identifying the visual monitoring hidden danger of the overhead transmission line are realized when the processor executes the instructions.
The above description of the apparatus according to the method embodiment may further include other implementations, and the implementation principle and the generated technical effects of the embodiment are the same as those of the foregoing method embodiment, and specific reference may be made to the description of the related method embodiment, which is not repeated herein.
Finally, it should be noted that: the above examples are only specific embodiments of the present invention, and are not intended to limit the scope of the present invention, but it should be understood by those skilled in the art that the present invention is not limited thereto, and that the present invention is described in detail with reference to the foregoing examples: any person skilled in the art may modify or easily conceive of the technical solution described in the foregoing embodiments, or perform equivalent substitution of some of the technical features, while remaining within the technical scope of the present disclosure; such modifications, changes or substitutions do not depart from the spirit and scope of the corresponding technical solutions. Are intended to be encompassed within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims (10)

1. The method for identifying the hidden trouble of the visual monitoring of the overhead transmission line is characterized by comprising the following steps of:
s1: acquiring an image which is captured at regular time through a monitoring device, detecting hidden danger of the image through a trained hidden danger identification model, judging whether hidden danger exists or not, and judging the type of the hidden danger; wherein the hidden danger types comprise construction machinery, lead foreign matters and smoke forest fires;
s2: when the hidden danger type is judged to be construction machinery or smoke mountain fire, S3 is executed; when the hidden danger type is judged to be the foreign matter of the lead, judging the alarm level as critical alarm, and executing S4;
s3: judging the alarm level as a general alarm, a key alarm or a critical alarm according to the movement trend of the construction machine or the smoke forest fire and the position relation between the alarm position of the construction machine or the smoke forest fire and the set protection area of the overhead transmission line;
s4: for general alarms, the process is not performed, and the next patrol period is waited to be executed from S1; for important alarms, carrying out timing intensive snapshot through a monitoring device, and outputting important alarm information and images of the timing intensive snapshot; and for critical alarming, adjusting the position and focal length of the monitoring device, collecting video streams of the alarming position, analyzing the video streams, and outputting critical alarming information and analysis results of the video streams.
2. The method for identifying hidden dangers in visualized monitoring of an overhead transmission line according to claim 1, characterized in that the protection area of the overhead transmission line is set by the following process:
acquiring target images of a wire and a tower of an overhead transmission line, and detecting the tower and the wire of the target images through a tower detection model and a wire detection model respectively to obtain a tower position and a wire position;
determining a main tower according to the relative position information of the wire position and the tower position, determining the outermost wires at two sides of the main tower by using the wire position, and determining the area between the main tower and the outermost wires as an initial protection area;
and (3) expanding the initial protection area according to the voltage class of the overhead transmission line to obtain the protection area of the overhead transmission line.
3. The method for identifying hidden dangers in visualized monitoring of an overhead transmission line according to claim 2, characterized in that the initial protection area is subjected to expansion by the following process:
setting an expansion coefficient K and calculating an expansion distance R;
wherein, (r=x R -X L )*K/2，X R and XL The X coordinates of the right edge point and the X coordinates of the left edge point of the initial protection area are respectively;
for each contour point P of the initial protection zone i When the outline point P i When the points are convex, concave or flat, the outline point Q after expansion is calculated by the following formulas (1), (2) and (3) i ；
wherein ,
P i-1 and Pi+1 Respectively the contour points P i And a preceding contour point and a following contour point, θ i Is a line segment P i-1 P i And P i+1 P i Included angle e i Is P i-1 Pointing to P i Unit vector of e i+1 Is P i Pointing to P i+1 Is a unit vector of (a).
4. The method for identifying hidden dangers in visualized monitoring of an overhead transmission line according to claim 1, wherein the step S3 comprises:
s31: calculating the intersection ratio IOU1 of a detection frame of the alarm position of the smoke mountain fire of the front image and the rear image, comparing the intersection ratio IOU1 with a set threshold T1, judging the alarm level as important alarm if the IOU1 is smaller than T1, otherwise executing S32;
calculating the intersection ratio IOU2 of a detection frame of the warning position of the construction machine of the front image and the rear image, comparing the intersection ratio IOU2 with a set threshold T2, judging the warning level as a general warning if the IOU2 is larger than T2, otherwise executing S32;
s32: judging whether an intersection exists between a detection frame of the warning position of the construction machine or the smoke mountain fire and a protection area of the overhead transmission line, if not, judging the warning grade as a general warning, and if so, executing S33;
S33: if the alarm position of the construction machine or the smoke forest fire is located outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is larger than the first distance, judging the alarm grade as a general alarm;
if the alarm position of the construction machine or the smoke forest fire is located outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is between the second distance and the first distance, judging the alarm level as a key alarm;
if the alarm position of the construction machine or the smoke forest fire is located in the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is smaller than the second distance, judging the alarm level as critical alarm;
wherein the second distance is less than the first distance.
5. The method for identifying hidden dangers in visual monitoring of an overhead transmission line according to any one of claims 1 to 4, characterized in that the video stream is analyzed by the following process:
and detecting hidden danger of each frame of image of the video stream through a video detection model, continuously analyzing the detected hidden danger through a target tracking model, and calculating the motion state of the construction machine or the diffusion trend of smoke mountain fire through a trend judging method.
6. The method for identifying hidden dangers in visualized monitoring of an overhead transmission line according to claim 5, wherein the position of the monitoring device is adjusted by the following process:
calculating an X-direction adjustment scale angle_x and a Y-direction adjustment scale angle_y of the monitoring device, and adjusting the position of the monitoring device according to the angle_x and the angle_y;
Angle_x＝(bbox_center_x-image_center_x)*FOV_width/width
Angle_y＝(bbox_center_y-image_center_y)*FOV_height/height
wherein bbox_center_x and bbox_center_y are the X and Y coordinates of the center of the alert position, image_center_x and image_center_y are the X and Y coordinates of the center of the image, fov_width and fov_height are the width and height views of the monitor, respectively, and width and height are the width and height of the image, respectively;
the focal length of the monitoring device is adjusted through the following process:
calculating a New focal length New_focal of the monitoring device, and adjusting the focal length of the monitoring device according to the New focal length New_focal;
New_focal＝current_focal+bbox_width/width*scale
the current_focal is the current focal length of the monitor, bbox_width is the width of the monitor, and scale is the set scaling factor.
7. Visual prison of overhead transmission line claps hidden danger recognition device, its characterized in that, the device includes:
the hidden danger detection module is used for acquiring images which are captured at regular time through the monitoring device, detecting hidden danger of the images through the trained hidden danger identification model, judging whether hidden danger exists or not and judging the type of the hidden danger; wherein the hidden danger types comprise construction machinery, lead foreign matters and smoke forest fires;
The first level judging module is used for executing the second level judging module when the hidden danger type is judged to be construction machinery or smoke mountain fire; when the hidden danger type is judged to be the foreign matter of the lead, judging the alarm level as critical alarm, and executing an alarm processing module;
the second level judging module is used for judging the alarm level as a general alarm, a key alarm or a critical alarm according to the movement trend of the construction machinery or the smoke mountain fire and the position relation between the alarm position of the construction machinery or the smoke mountain fire and the set protection area of the overhead transmission line;
the alarm processing module is used for processing general alarms, and waiting for the execution of the next inspection period from the hidden danger detection module; for important alarms, carrying out timing intensive snapshot through a monitoring device, and outputting important alarm information and images of the timing intensive snapshot; and for critical alarming, adjusting the position and focal length of the monitoring device, collecting video streams of the alarming position, analyzing the video streams, and outputting critical alarming information and analysis results of the video streams.
8. The overhead transmission line visualization monitoring hidden trouble recognition device of claim 7, wherein the second level judgment module comprises:
The motion trend judging unit is used for calculating the intersection ratio IOU1 of the detection frames of the smoke mountain fire alarm positions of the front image and the rear image, comparing the intersection ratio IOU1 with a set threshold T1, judging the alarm level as a key alarm if the IOU1 is smaller than T1, and executing the qualitative analysis unit if the alarm level is not smaller than T1;
calculating the intersection ratio IOU2 of a detection frame of the warning position of the construction machine of the front image and the rear image, comparing the intersection ratio IOU2 with a set threshold T2, judging the warning grade as a general warning if the IOU2 is larger than the threshold T2, and executing a qualitative analysis unit if the IOU2 is not larger than the threshold T2;
the qualitative analysis unit is used for judging whether an intersection exists between a detection frame of the warning position of construction machinery or smoke mountain fire and a protection area of the overhead transmission line, if not, judging the warning grade as a general warning, and if so, executing the quantitative analysis unit;
the quantitative analysis unit is used for judging the alarm grade as a general alarm if the alarm position of the construction machine or the smoke mountain fire is located outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke mountain fire and the wire position is larger than a first distance;
if the alarm position of the construction machine or the smoke forest fire is located outside the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is between the second distance and the first distance, judging the alarm level as a key alarm;
If the alarm position of the construction machine or the smoke forest fire is located in the protection area of the overhead transmission line or the nearest distance between the alarm position of the construction machine or the smoke forest fire and the wire position is smaller than the second distance, judging the alarm level as critical alarm;
wherein the second distance is less than the first distance.
9. A computer readable storage medium for identifying visual monitoring hidden trouble of overhead transmission line, comprising a memory for storing instructions executable by a processor, which when executed by the processor, implement steps comprising the method for identifying visual monitoring hidden trouble of overhead transmission line of any one of claims 1-6.
10. An apparatus for identifying potential overhead transmission line visual inspection hazards, comprising at least one processor and a memory storing computer executable instructions, wherein the processor, when executing the instructions, performs the steps of the method for identifying potential overhead transmission line visual inspection hazards of any one of claims 1-6.
CN202310548207.1A 2023-05-15 2023-05-15 Visual monitoring hidden trouble identification method and device for overhead transmission line Pending CN116797967A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310548207.1A CN116797967A (en) 2023-05-15 2023-05-15 Visual monitoring hidden trouble identification method and device for overhead transmission line

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310548207.1A CN116797967A (en) 2023-05-15 2023-05-15 Visual monitoring hidden trouble identification method and device for overhead transmission line

Publications (1)

Publication Number Publication Date
CN116797967A true CN116797967A (en) 2023-09-22

Family

ID=88048929

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310548207.1A Pending CN116797967A (en) 2023-05-15 2023-05-15 Visual monitoring hidden trouble identification method and device for overhead transmission line

Country Status (1)

Country Link
CN (1) CN116797967A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118097565A (en) * 2024-04-19 2024-05-28 国网上海市电力公司 Method for judging invalid hidden trouble of power transmission channel based on monocular vision

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN118097565A (en) * 2024-04-19 2024-05-28 国网上海市电力公司 Method for judging invalid hidden trouble of power transmission channel based on monocular vision
CN118097565B (en) * 2024-04-19 2024-07-30 国网上海市电力公司 Method for judging invalid hidden trouble of power transmission channel based on monocular vision

Similar Documents

Publication Publication Date Title
CN111783744A (en) Operation site safety protection detection method and device
US20230005176A1 (en) Throwing position acquisition method and apparatus, computer device and storage medium
CN101908142A (en) Feature analysis-based video flame detecting method
KR20100119476A (en) An outomatic sensing system for traffic accident and method thereof
CN105976365A (en) Nocturnal fire disaster video detection method
CN111461078B (en) Fishing preventing and monitoring method based on computer vision technology
CN116743970B (en) Intelligent management platform with video AI early warning analysis
CN111178424A (en) Petrochemical production site safety compliance real-time detection system and method
Lee et al. Real-time fire detection using camera sequence image in tunnel environment
CN116797967A (en) Visual monitoring hidden trouble identification method and device for overhead transmission line
CN110956652A (en) Early warning method for transformer substation personnel crossing line
CN115457331A (en) Intelligent inspection method and system for construction site
CN110913209A (en) Camera shielding detection method and device, electronic equipment and monitoring system
JP4302801B2 (en) Object tracking method and object tracking apparatus
JP6978986B2 (en) Warning system, warning control device and warning method
CN114699702B (en) Fire fighting equipment detection method and related device
CN116246206A (en) Coal conveying system coal overflow identification system based on marker edge detection
CN115115988A (en) Personnel falling risk detection method and system and storage medium
CN113378728A (en) Monitoring method, system, electronic equipment and computer readable storage medium
JP5309069B2 (en) Smoke detector
CN114445779A (en) Method, system, device and medium for determining safety state of to-be-detected area
JP5329466B2 (en) Smoke detector
CN106803322B (en) A kind of anti-limited region dividing method
CN113298048B (en) Forest fire detection and early warning system based on computer vision
CN110309838A (en) Video detection area contour of object based on exponential transform detects preprocess method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination