CN-121074796-B - Method and device for monitoring construction safety of pole tower

CN121074796BCN 121074796 BCN121074796 BCN 121074796BCN-121074796-B

Abstract

The application relates to the technical field of electric power and computer vision, in particular to a method and a device for monitoring the construction safety of a pole and a tower. And identifying construction stages, construction atmosphere, dangerous construction behaviors and personnel construction experience based on the video stream. And correcting the initial risk coefficient of the dangerous construction behavior by using the second correction coefficient of the construction stage, construction atmosphere, personnel experience and weather for the video stream with the dangerous construction behavior to obtain a first risk coefficient, and arranging the first risk coefficients in descending order. And correcting a second risk coefficient of the initial risk coefficient of the construction stage by using a fourth correction coefficient of construction atmosphere, personnel experience and weather for the video stream without dangerous behaviors, wherein the video stream without dangerous behaviors is arranged in the descending order, and the video of the dangerous behaviors is in front. The application can more accurately evaluate the actual dangerous degree of the current dangerous construction behavior, more accurately evaluate the potential dangerous degree of the current tower construction and greatly shorten the response link from risk occurrence to intervention.

Inventors

JIANG SHUCHU
YE HAIHONG
HAO GUIQIN
LIANG GUORUI
CAO TIANSHENG
WU LINLONG
TANG YONGNING
DENG ZHIMIN
LI NING

Assignees

广州电力建设有限公司

Dates

Publication Date: 20260512
Application Date: 20250902

Claims (7)

1. A method for monitoring construction safety of a pole tower, which is characterized by comprising the following steps: Acquiring video streams of construction sites of the pole towers of all the machine positions in real time; Respectively identifying the current construction stage, the current construction atmosphere, the current dangerous construction behavior and the tower construction experience of current constructors based on video streams of the pole and tower construction sites of all machine positions; Synchronously acquiring the current weather of each machine position pole tower construction site; for a video stream with dangerous construction behaviors, acquiring a second correction coefficient of the current dangerous construction behaviors from a preset first association table, wherein the second correction coefficient is used for correcting an initial dangerous coefficient of the current dangerous construction behaviors to obtain a first dangerous coefficient; arranging all the machine position monitoring videos with dangerous construction behaviors in descending order according to the first dangerous coefficient; for the video stream without dangerous construction behaviors, acquiring a fourth correction coefficient of the current construction stage from a preset second association table for correcting the initial dangerous coefficient of the current construction stage to obtain a second dangerous coefficient, wherein the fourth correction coefficient comprises a current construction atmosphere, a tower construction experience of a current constructor and a third correction coefficient of a current weather; arranging all the machine position monitoring videos without dangerous construction behaviors according to a second dangerous coefficient descending order; Outputting all sequenced monitoring videos to a background monitoring screen for display, wherein the monitoring videos with dangerous construction behaviors are arranged before the monitoring videos without dangerous construction behaviors; The step of identifying the current construction stage based on the video stream of the pole and tower construction site comprises the following steps: preprocessing the video stream of the tower construction site to obtain a preprocessed tower video stream; extracting key frames from the preprocessed tower video stream, wherein the key frames are extracted by detecting motion changes among the video frames through an optical flow method, and screening frames with obvious motion changes as key frames; Inputting the extracted key frames into a pre-trained space-time feature extraction model to obtain space-time features, wherein the space-time feature extraction model comprises a cascaded 3D convolutional neural network and a two-way long-short-term memory network; Fusing the space-time features and the IoT data of the shaft tower construction site in the channel dimension through a transducer cross attention mechanism to obtain fused features; Inputting the fused feature vectors into a Softmax classifier, and outputting the category of the tower construction stage at the current moment, wherein the category at least comprises foundation excavation and grounding groove construction, pre-splicing and transporting of the tower material ground, assembling of tower legs and bottom sections, high-level lifting of tower body fragments and sections, integral lifting of cross arms, permanent wire tensioning and pole holding dismantling, bolt tightening and clearing and defect elimination.
2. The method for monitoring the construction safety of a pole tower according to claim 1, wherein the step of obtaining the second correction coefficient of the current dangerous construction behavior from the preset first correlation table for correcting the initial dangerous coefficient of the current dangerous construction behavior to obtain the first dangerous coefficient comprises the steps of: Correcting the initial dangerous coefficient of the current dangerous construction behavior according to the following model; ; Wherein, the As a first risk factor, the first risk factor, As an initial risk factor for the current dangerous construction behavior, As the first correction factor for the current construction stage, Is the first correction coefficient of the current construction atmosphere, A first correction factor for the tower construction experience of the current constructor, Is the first correction factor for the current weather.
3. The method for monitoring the construction safety of a tower according to claim 1, wherein the step of obtaining the fourth correction coefficient of the current construction stage from the preset second association table for correcting the initial risk coefficient of the current construction stage to obtain the second risk coefficient comprises: correcting the initial risk coefficient of the current construction stage according to the following model to obtain a second risk coefficient; ; Wherein, the As a second risk factor, the first risk factor, As an initial risk factor for the current construction stage, A third correction coefficient for the current construction atmosphere, A third correction factor for the tower construction experience of the current constructor, And a third correction factor for the current weather.
4. The method for monitoring the construction safety of the pole tower according to claim 1, wherein the first correction coefficient is calculated by the following method: collecting first construction history data of a pole tower, wherein the first construction history data of the pole tower comprises a dangerous construction behavior identifier, an initial dangerous coefficient of dangerous construction behavior, a construction stage, a construction atmosphere, pole tower construction experience of constructors, weather and corresponding first actual risk results; Preprocessing the first construction history data of the pole tower, including quantifying the first actual risk result, and grouping according to dangerous construction behavior types, construction stages, construction atmosphere, pole tower construction experience of constructors and gear combinations of weather; Constructing a first multiple regression model, wherein the quantized first actual risk result is taken as a dependent variable, and the first correction coefficient of a construction stage, the first correction coefficient of a construction atmosphere, the first correction coefficient of the tower construction experience of constructors and the first correction coefficient of weather are taken as independent variables; dividing the preprocessed first construction history data of the tower into a training set and a testing set, training the first multiple regression model by using the training set, and solving each first correction coefficient which enables the sum of squares of the predicted value of the first multiple regression model and the quantized first actual risk result residual error to be minimum by using a least square method; And verifying the trained first multiple regression model by using a test set, if the prediction error is within a preset range, determining each obtained first correction coefficient as a final first correction coefficient, and if the prediction error is not within the preset range, adjusting the data sample until each first correction coefficient meeting the error requirement is obtained.
5. The method for monitoring the construction safety of the pole tower according to claim 1, wherein the third correction coefficient is calculated by the following method: Collecting second construction history data of the pole tower, wherein the second construction history data of the pole tower comprise a construction stage, an initial danger coefficient of the construction stage, a construction atmosphere, pole tower construction experience of constructors, weather and corresponding second actual risk results; Preprocessing the second construction history data of the tower, including quantifying the second actual risk result, and grouping according to construction stages, construction atmosphere, tower construction experience of constructors and gear combination of weather; Constructing a second multiple regression model, wherein the second quantized actual risk result is taken as a dependent variable, and a third correction coefficient of construction atmosphere, a third correction coefficient of tower construction experience of constructors and a third correction coefficient of weather are taken as independent variables; dividing the preprocessed second construction history data of the tower into a training set and a testing set, training the second multiple regression model by using the training set, and solving each third correction coefficient which enables the sum of squares of the predicted value of the second multiple regression model and the quantized second actual risk result residual error to be minimum by using a least square method; And verifying the trained second multiple regression model by using a test set, if the prediction error is within a preset range, determining each obtained third correction coefficient as a final third correction coefficient, and if the prediction error is not within the preset range, adjusting the data sample until each third correction coefficient meeting the error requirement is obtained.
6. The method for monitoring the construction safety of a pole and a tower according to claim 1, wherein the step of identifying the current construction atmosphere based on the video stream of the construction site of the pole and the tower comprises Calculating the score of the current construction atmosphere according to the following model; ; ; ; ; Wherein X is the current construction atmosphere score, Scoring the construction state of the current ith constructor and jth constructor, m is the number of identified constructors, n is the number of evaluation times, Is the probability value of the sigma emotion corresponding to the j constructor, A sigma feature value extracted for a j-th constructor face, wherein sigma = 1,2,..s, an identification dimension corresponding to s emotions, And Is a parameter for adjusting the linear relation between emotion characteristics and construction emotion, The weight value of the sigma-th facial feature of the j-th constructor; The mapping function is used for mapping the result to a preset construction state score; and mapping the score of the current construction atmosphere to a construction atmosphere type.
7. A tower construction safety monitoring device, the device comprising: the video stream acquisition module is used for acquiring video streams of the construction sites of the machine position towers in real time; The identifying module is used for respectively identifying the current construction stage, the current construction atmosphere, the current dangerous construction behavior and the tower construction experience of the current constructor based on the video stream of the construction site of each machine position tower; The weather acquisition module is used for synchronously acquiring the current weather of each machine position pole tower construction site; The system comprises a first correction module, a first correction module and a second correction module, wherein the first correction module is used for acquiring a second correction coefficient of a current dangerous construction behavior from a preset first correlation table for correcting an initial dangerous coefficient of the current dangerous construction behavior to obtain a first dangerous coefficient for a video stream with the dangerous construction behavior, and the second correction coefficient comprises a current construction stage, a current construction atmosphere, a tower construction experience of a current constructor and a first correction coefficient of a current weather; The first arrangement module is used for arranging the monitoring videos of all the machine positions with dangerous construction behaviors in descending order according to the first dangerous coefficient; The second correction module is used for acquiring a fourth correction coefficient of the current construction stage from a preset second association table for correcting the initial risk coefficient of the current construction stage to obtain a second risk coefficient for the video stream without dangerous construction behaviors, wherein the fourth correction coefficient comprises a current construction atmosphere, a tower construction experience of a current constructor and a third correction coefficient of a current weather; the second arrangement module is used for arranging the monitoring videos of all the machine positions without dangerous construction behaviors in descending order according to a second dangerous coefficient; The system comprises an output module, a background monitoring screen and a monitoring module, wherein the output module is used for outputting all sequenced monitoring videos to the background monitoring screen for display, and the monitoring videos with dangerous construction behaviors are arranged before the monitoring videos without dangerous construction behaviors; The video stream identification based on the pole and tower construction site comprises the following steps: preprocessing the video stream of the tower construction site to obtain a preprocessed tower video stream; extracting key frames from the preprocessed tower video stream, wherein the key frames are extracted by detecting motion changes among the video frames through an optical flow method, and screening frames with obvious motion changes as key frames; Inputting the extracted key frames into a pre-trained space-time feature extraction model to obtain space-time features, wherein the space-time feature extraction model comprises a cascaded 3D convolutional neural network and a two-way long-short-term memory network; Fusing the space-time features and the IoT data of the shaft tower construction site in the channel dimension through a transducer cross attention mechanism to obtain fused features; Inputting the fused feature vectors into a Softmax classifier, and outputting the category of the tower construction stage at the current moment, wherein the category at least comprises foundation excavation and grounding groove construction, pre-splicing and transporting of the tower material ground, assembling of tower legs and bottom sections, high-level lifting of tower body fragments and sections, integral lifting of cross arms, permanent wire tensioning and pole holding dismantling, bolt tightening and clearing and defect elimination.

Description

Method and device for monitoring construction safety of pole tower Technical Field The application relates to the technical field of electric power and computer vision, in particular to a method and a device for monitoring the construction safety of a pole tower. Background In order to monitor the construction safety of the tower, at present, a camera erected on the ground or an unmanned aerial vehicle carrying the camera is generally used for safety monitoring, and monitoring videos are synchronously sent to a background display for display. The monitoring videos of different machine positions are displayed on a background display according to a fixed arrangement mode. This fixed ordering results in a failure of the monitoring personnel to quickly discover dangerous behavior and potential construction risks occurring during the construction of the pole and tower, resulting in failure to shorten the response link from risk occurrence to intervention. In addition, when different dangerous behaviors exist in different machine positions, the monitoring personnel are difficult to quickly identify the actual dangerous degree among the different dangerous behaviors, and when the dangerous behaviors do not occur in the different machine positions and are in different construction stages, the monitoring personnel are difficult to quickly identify the potential construction dangerous degree in the different construction stages. And because different monitoring personnel have differences on actual dangerous understanding of different behaviors, the objective and quantitative dangerous degree evaluation standard is lacked, and resource mismatch is caused, for example, excessive intervention resources are input on actual low-risk behaviors or insufficient importance is paid on high-risk behaviors, and the supervision efficiency is low. Disclosure of Invention The application aims to provide a method and a device for monitoring the construction safety of a pole tower, aiming at solving the technical problems. In a first aspect, an embodiment of the present application provides a method for monitoring construction safety of a tower, including: Acquiring video streams of construction sites of the pole towers of all the machine positions in real time; Respectively identifying the current construction stage, the current construction atmosphere, the current dangerous construction behavior and the tower construction experience of current constructors based on video streams of the pole and tower construction sites of all machine positions; Synchronously acquiring the current weather of each machine position pole tower construction site; for a video stream with dangerous construction behaviors, acquiring a second correction coefficient of the current dangerous construction behaviors from a preset first association table, wherein the second correction coefficient is used for correcting an initial dangerous coefficient of the current dangerous construction behaviors to obtain a first dangerous coefficient; arranging all the machine position monitoring videos with dangerous construction behaviors in descending order according to the first dangerous coefficient; for the video stream without dangerous construction behaviors, acquiring a fourth correction coefficient of the current construction stage from a preset second association table for correcting the initial dangerous coefficient of the current construction stage to obtain a second dangerous coefficient, wherein the fourth correction coefficient comprises a current construction atmosphere, a tower construction experience of a current constructor and a third correction coefficient of a current weather; arranging all the machine position monitoring videos without dangerous construction behaviors according to a second dangerous coefficient descending order; And outputting all the sequenced monitoring videos to a background monitoring screen for display, wherein the monitoring videos with dangerous construction behaviors are arranged before the monitoring videos without dangerous construction behaviors. Further, the step of obtaining the first risk coefficient by obtaining the second correction coefficient of the current dangerous construction behavior from the preset first association table to correct the initial risk coefficient of the current dangerous construction behavior includes: Correcting the initial dangerous coefficient of the current dangerous construction behavior according to the following model; R1＝a1×a2×a3×a4×R0; Wherein, R 1 is a first dangerous coefficient, R 0 is an initial dangerous coefficient of current dangerous construction behavior, a 1 is a first correction coefficient of current construction stage, a 2 is a first correction coefficient of current construction atmosphere, a 3 is a first correction coefficient of tower construction experience of current constructor, and a 4 is a first correction coefficient of current weather. Further, the step of obtainin