Automatic monitoring management method and system for super-dangerous large engineering and computer equipment

Abstract

The invention relates to the field of image data identification processing, and provides an automatic monitoring management method and system for a super-dangerous large project, and computer equipment, and the method comprises the steps: firstly carrying out the preprocessing of a monitoring image of a construction site, so as to adapt to the complex construction environment of the super-dangerous large project, guarantee the precise monitoring under the conditions of insufficient brightness or smoke and haze, and improve the construction efficiency. A dual-path context awareness network module is designed, the quality of basic image features is improved, the overall monitoring precision is improved, meanwhile, the wide-area relation and the context relation between the image features are guided and enhanced, precise monitoring of multiple complex targets in the wide-area range is achieved, a management judgment module is designed, and the management judgment module is used for managing the multiple targets in the wide-area range. According to the method, the monitoring adjustability is realized, the problem of false alarm or false alarm caused by different construction standards and construction conditions among various super-dangerous large projects is avoided, the automatic monitoring management of the super-dangerous large projects is realized, and the monitoring precision under the condition of a wide monitoring range is greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of image data recognition and processing, and particularly to an automated monitoring and management method, system, and computer device for extremely dangerous and large-scale projects. Background Art

[0002] With the rapid development of engineering technology, the management of engineering projects has gradually received attention. Especially for extremely dangerous and large-scale projects, due to their large project scale, numerous involved personnel, and extremely high risks during the construction process, they have become extremely important and difficult targets in project management.

[0003] In the prior art, the monitoring and management of extremely dangerous and large-scale projects mainly rely on on-site inspections and records by construction workers and safety supervisors. However, in the face of a construction environment with a vast scope and extremely high risks, the traditional manual management method is still difficult to conduct comprehensive inspections, and real-time management cannot be carried out throughout the entire construction process. It can only exist in the form of sampling inspections or multiple inspections in different areas. The labor cost required to achieve real-time management is extremely high, further increasing the management difficulty of extremely dangerous and large-scale projects. However, the monitoring accuracy and monitoring scope of existing automated monitoring and management methods cannot meet the monitoring and management requirements of extremely dangerous and large-scale projects with large project scales and numerous involved personnel.

[0004] Therefore, how to design an automated monitoring and management method to meet the automated monitoring and management requirements of extremely dangerous and large-scale projects has become an urgent problem to be solved. Summary of the Invention

[0005] Based on this, the purpose of the present invention is to provide an automated monitoring and management method, system, and computer device for extremely dangerous and large-scale projects. By first preprocessing the monitoring images at the construction site to adapt to the complex construction environment of extremely dangerous and large-scale projects and ensuring precise monitoring in case of insufficient light or smoke, haze, and fog, and then designing a dual-path scenario perception network module, including a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module, the quality of basic image features is improved to enhance the overall monitoring accuracy. At the same time, the wide-area relationship and context relationship between image features are guided and enhanced to achieve precise monitoring of complex multi-targets in a wide area. A management judgment module is also designed to achieve the adjustability of monitoring, avoiding false alarms or misreports caused by different construction standards and construction conditions among various extremely dangerous and large-scale projects, and further improving the monitoring accuracy. The present invention realizes the automated monitoring and management of extremely dangerous and large-scale projects, greatly improving the monitoring accuracy in the case of a wide-area monitoring scope.

[0006] An automated management method for extremely dangerous and large-scale projects proposed by the present invention includes: The acquisition module obtains the construction site data in real time. The construction site data includes construction site image data, construction site luminance data, and construction site haze degree data. The construction site image data is preprocessed and then transmitted to the dual-path scenario awareness network module; The dual-path scenario awareness network module includes a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module. The dual-path attention weighting sub-module includes a spatial attention mechanism path and a channel attention mechanism path. The basic feature extraction sub-module obtains the basic image features of the construction site image data. The dual-path attention weighting sub-module performs attention weighting processing on the basic image features according to the attention mechanism to obtain the attention-weighted image features corresponding to the attention mechanism. The attention-weighted image features and the basic image features are input into the multi-head linear self-attention sub-module to obtain an attention-enhanced image, which is then transmitted to the management judgment module; The management judgment module judges the construction standards of the attention-enhanced image according to the preset construction database. The preset construction database includes construction drawing data, construction personnel data, and construction safety standard data. If it is determined that the current construction state does not meet the construction standards, a construction problem alarm will be issued.

[0007] In summary, according to the above-mentioned automated management method for extremely dangerous and major projects, by first preprocessing the monitoring images at the construction site to adapt to the complex construction environment of extremely dangerous and major projects and ensuring accurate monitoring in the case of insufficient light or dust, haze, etc., and then by designing a dual-path scenario awareness network module, including a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module, the quality of the basic image features is improved to enhance the overall monitoring accuracy. At the same time, the wide-area relationship and context relationship between image features are guided and enhanced, realizing accurate monitoring of complex multi-targets in a wide area. A management judgment module is also designed to achieve the adjustability of monitoring, avoiding false alarms or misreports caused by different construction standards and construction conditions among various extremely dangerous and major projects, and further improving the monitoring accuracy. The present invention realizes the automated monitoring management of extremely dangerous and major projects, greatly improving the monitoring accuracy in the case of a wide-area monitoring range.Specifically, the acquisition module obtains construction site data in real time. The construction site data includes construction site image data, construction site luminance data, and construction site haze data. The construction site image data is preprocessed and then transmitted to the dual-path scenario awareness network module, eliminating the influence of dim environments and smoky and hazy environments, and avoiding the decrease in monitoring accuracy caused by the complex construction environment of ultra-high-risk and large-scale projects. The dual-path scenario awareness network module includes a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module. The dual-path attention weighting sub-module includes a spatial attention mechanism path and a channel attention mechanism path. The basic feature extraction sub-module obtains the basic image features of the construction site image data, and performs targeted extraction of basic features from the aspects of global wide-area scale, multi-local detail scale, and deep scale, improving the quality of the basic features, and thus improving the overall monitoring accuracy. The dual-path attention weighting sub-module performs attention weighting processing on the basic image features according to the attention mechanism to obtain attention-weighted image features corresponding to the attention mechanism, enhancing the feature quality of a single target person and target area in the case of a large range of multiple targets, and thus improving the monitoring accuracy of a single target. The attention-weighted image features and the basic image features are input into the multi-head linear self-attention sub-module to obtain attention-enhanced images and transmitted to the management judgment module, enhancing the feature relationship of multiple target persons and target areas in the wide-area range, and thus improving the monitoring accuracy in the wide-area range. The management judgment module makes a construction standard judgment on the attention-enhanced image according to the preset construction database. The preset construction database includes construction drawing data, construction personnel data, and construction safety standard data. If it is determined that the current construction state does not conform to the construction standard, a construction problem alarm is issued, realizing the adjustability of monitoring and avoiding false alarms or misreports caused by different construction standards and construction conditions among multiple ultra-high-risk and large-scale projects, further improving the monitoring accuracy. The present invention realizes the automated monitoring management of ultra-high-risk and large-scale projects, greatly improving the monitoring accuracy in the case of a wide-area monitoring range.

[0008] Further, the acquisition module obtains construction site data in real time. The construction site data includes construction site image data, construction site luminance data, and construction site haze data. The steps of preprocessing the construction site image data include: The acquisition module obtains construction site data in real time and performs image enhancement preprocessing. The construction site data includes construction site image data, construction site luminance data, and construction site haze data; Judge whether the brightness of the construction site is less than the preset brightness threshold according to the brightness data of the construction site. If it is determined that the brightness of the construction site is less than the preset brightness threshold, perform brightness enhancement preprocessing on the image data of the construction site, extract the pixel values of the red, green, and blue channels in the image data of the construction site, and perform weighted grayscale processing on the image data of the construction site according to the following formula: where, represents the weighted grayscale value of each pixel point, a represents the total number of pixel points, represents the maximum value among the pixel values of the three channels of each pixel point of the grayscale weight value, represents the minimum value among the pixel values of the three channels of each pixel point of the grayscale weight value, R, G, and B respectively represent the pixel values of the red, green, and blue channels of each pixel point. After linearly normalizing the image data of the construction site, perform image grayscale averaging processing according to the following formula: where, represents the grayscale average value of the pixel point, and use the grayscale average value as the image grayscale value of the image data of the construction site; Judge whether the haze degree of the construction site is greater than the preset haze degree threshold according to the haze degree data of the construction site. If it is determined that the haze degree of the construction site is greater than the preset haze degree threshold, perform haze removal preprocessing on the image data of the construction site, estimate the atmospheric light value A according to the image grayscale value of the image data of the construction site, and perform calculation and guided filtering processing according to the estimated atmospheric light value A to obtain the transmittance , and obtain a preprocessed image according to the atmospheric light value A and the transmittance . The formula for obtaining the preprocessed image is as follows: where, represents the preprocessed image, represents the unprocessed image, represents the lower threshold of the transmittance, and max(·) represents obtaining the maximum value.

[0009] Furthermore, the steps for the basic feature extraction sub-module to obtain the basic image features of the image data of the construction site include: Input the preprocessed construction site image data into the basic feature extraction sub-module. The basic feature extraction sub-module includes a first convolutional layer, a second convolutional layer, and a third convolutional layer. The first convolutional layer includes two identical convolutional branches, and each convolutional branch includes 2 5X5 convolutional blocks. The first convolutional layer obtains global basic features. The second convolutional layer includes a bottleneck structure branch, an inverted bottleneck structure branch, and a residual branch. The bottleneck structure branch includes two 1X1 convolutional blocks and one 3X3 convolutional block. The inverted bottleneck structure branch includes one 1X1 convolutional block and two 3X3 convolutional blocks. The second convolutional layer obtains depth basic features. The third convolutional layer includes a multi-scale convolutional branch and a residual branch. The multi-scale convolutional branch includes one 1X1 convolutional block, one 2X2 convolutional block, and one 3X3 convolutional block. The third convolutional layer obtains multi-scale basic features. Fuse the global basic features, the depth basic features, and the multi-scale basic features to obtain basic image features.

[0010] Further, the steps of the dual-path attention weighting sub-module performing attention weighting processing on the basic image features according to the attention mechanism to obtain the attention-weighted image features corresponding to the attention mechanism include: Input the basic image features into the spatial attention mechanism path of the dual-path attention weighting sub-module, and perform spatial attention weighting processing on the basic image features according to the spatial attention mechanism. The spatial attention mechanism path of the dual-path attention weighting sub-module includes a global average pooling layer, a global maximum pooling layer, one 3X3 convolutional block, one 5X5 convolutional block, and one 7X7 convolutional block. The formula for the spatial attention weighting processing is as follows: Wherein, represents the basic image features, represents the spatially attention-weighted image features, represents the activation function, represents the convolutional block, represents concatenation, represents the global maximum pooling layer, represents the global average pooling layer; To obtain the spatially attention-weighted image features.

[0011] Further, the steps of the dual-path attention weighting sub-module performing attention weighting processing on the basic image features according to the attention mechanism to obtain the attention-weighted image features corresponding to the attention mechanism further include: Input the basic image features into the channel attention mechanism path of the dual-path attention weighting sub-module, and perform channel attention weighting processing on the basic image features according to the channel attention mechanism. The channel attention mechanism path of the dual-path attention weighting sub-module includes a global average pooling layer, a global max pooling layer, and a multi-layer perceptron. The formula for the channel attention weighting processing is as follows: Among them, represents the basic image features, represents the channel attention weighted image features, represents the activation function, represents the multi-layer perceptron, represents the global max pooling layer, represents the global average pooling layer; To obtain the channel attention weighted image features.

[0012] Further, the step of inputting the attention weighted image features and the basic image features into the multi-head linear self-attention sub-module to obtain the attention enhanced image and transmitting it to the management judgment module includes: Input the spatial attention weighted image features, the channel attention weighted image features, and the basic image features into the linear self-attention sub-module, and respectively use them as the Q matrix, K matrix, and V matrix of the linear self-attention sub-module to obtain the attention enhanced image features. The mechanism of the linear self-attention sub-module is as follows: Among them, represents Q matrix weight, represents K matrix weight, represents V matrix weight, represents the depth convolution, represents Q matrix and K the dimension of the matrix, represents the transposed matrix; Transmit the obtained attention enhanced image to the management judgment module.

[0013] Further, the step of the management judgment module judging the construction standard according to the preset construction database for the attention enhanced image includes: After the management judgment module obtains the attention-enhanced image features, it determines whether the current construction of the ultra-high-risk and major project complies with the construction standards according to the data labels in the preset construction database. The construction data labels include the construction area label and the construction structure label in the construction drawing data, the construction personnel quantity label and the construction equipment label used by the construction personnel in the construction personnel data, and the construction dangerous area label and the construction personnel safety equipment label in the construction safety standard data. The construction area label, the construction structure label, the construction personnel quantity label, and the construction equipment label used by the construction personnel are safety labels, and the construction dangerous area label and the construction personnel safety equipment label are dangerous labels. If a safety label is missing or a dangerous label exists in the attention-enhanced image, an alarm for construction problems of the ultra-high-risk and major project is given.

[0014] An ultra-high-risk and major project automatic management system proposed by the present invention includes: An acquisition module, which is used to acquire the construction site data in real time. The construction site data includes the construction site image data, the construction site brightness data, and the construction site haze data. The construction site image data is preprocessed and then transmitted to the dual-path scenario perception network module. The dual-path scenario perception network module, where the dual-path scenario perception network module includes a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module. The dual-path attention weighting sub-module includes a spatial attention mechanism path and a channel attention mechanism path. The basic feature extraction sub-module obtains the basic image features of the construction site image data. The dual-path attention weighting sub-module performs attention weighting processing on the basic image features according to the attention mechanism to obtain the attention-weighted image features corresponding to the attention mechanism. The attention-weighted image features and the basic image features are input into the multi-head linear self-attention sub-module to obtain an attention-enhanced image, which is then transmitted to the management judgment module. The management judgment module, which is used to judge the construction standards of the attention-enhanced image according to the preset construction database. The preset construction database includes construction drawing data, construction personnel data, and construction safety standard data. If it is determined that the current construction state does not meet the construction standards, an alarm for construction problems is given.

[0015] The present invention also provides a storage medium, where the storage medium stores one or more programs, and when the programs are executed by a processor, the above-mentioned ultra-high-risk and major project automatic monitoring and management method is implemented.

[0016] The present invention also provides a computer device, which includes a memory and a processor, where: The memory is used to store a computer program; When the processor is used to execute the computer program stored in the memory, the above-mentioned ultra-dangerous and major project automatic monitoring and management method is implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 FIG. 1 is a flowchart of an ultra-dangerous and major project automatic monitoring and management method according to a first embodiment of the present invention; Figure 2 FIG. 2 is a flowchart of an ultra-dangerous and major project automatic monitoring and management method according to a second embodiment of the present invention; Figure 3 FIG. 3 is a schematic structural diagram of an ultra-dangerous and major project automatic monitoring and management system according to a third embodiment of the present invention.

[0018] The following specific embodiments will further illustrate the present invention in conjunction with the above-mentioned drawings. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] In order to facilitate the understanding of the present invention, the present invention will be described more comprehensively below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, these embodiments are provided so that the disclosure of the present invention is more thorough and comprehensive.

[0020] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly on the other element or there can also be a middle element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there may be a middle element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are only for the purpose of illustration.

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs. The terms used in the description of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0022] Please refer to Figure 1 , which shows a flowchart of an ultra-dangerous and major project automatic monitoring and management method according to a first embodiment of the present invention. This ultra-dangerous and major project automatic monitoring and management method includes steps S01 to S03, wherein: Step S01: The acquisition module continuously obtains construction site data in real time. The construction site data includes construction site image data, construction site light brightness data, and construction site haze data. The construction site image data is preprocessed and then transmitted to the dual-path scenario perception network module; It should be noted that the preprocessing in this embodiment includes image preprocessing for a dim construction environment and preprocessing for a construction environment with soot, dust, haze, etc. The process of image preprocessing for a dim construction environment is as follows: based on the light brightness data of the construction site, it is determined whether the light brightness of the construction site is less than a preset light brightness threshold. If it is determined that the light brightness of the construction site is less than the preset light brightness threshold, then the light brightness enhancement preprocessing is performed on the image data of the construction site, the pixel values of the red, green, and blue channels in the image data of the construction site are extracted, and the image data of the construction site is weighted grayscale processed according to the following formula: Wherein, represents the weighted grayscale value of each pixel point, a represents the total number of pixel points, represents the grayscale weight value of the maximum value among the pixel values of the three channels of each pixel point ; represents the grayscale weight value of the minimum value among the pixel values of the three channels of each pixel point ; R, G, and B respectively represent the pixel values of the red channel, green channel, and blue channel of each pixel point. After linearly normalizing the image data of the construction site, the image grayscale averaging process is performed according to the following formula: Wherein, represents the grayscale average value of the pixel point. Taking the grayscale average value as the image grayscale value of the image data of the construction site, the preprocessing process for a construction environment with soot, dust, haze, etc. is as follows: based on the haze degree data of the construction site, it is determined whether the haze degree of the construction site is greater than a preset haze degree threshold. If it is determined that the haze degree of the construction site is greater than the preset haze degree threshold, then the haze removal preprocessing is performed on the image data of the construction site, the atmospheric light value A is estimated according to the image grayscale value of the image data of the construction site, and calculation and guided filtering processing are performed based on the estimated atmospheric light value A to obtain the transmittance , and based on the atmospheric light value A and the transmittance a preprocessed image is obtained. The formula for obtaining the preprocessed image is as follows: Wherein, represents the preprocessed image, represents the unprocessed image, represents the lower threshold of the transmittance, and max(·) represents obtaining the maximum value.

[0023] Step S02: The dual-path scenario awareness network module includes a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module. The dual-channel attention weighting sub-module includes a spatial attention mechanism path and a channel attention mechanism path. The basic feature extraction sub-module obtains the basic image features of the construction site image data. The dual-channel attention weighting sub-module performs attention weighting processing on the basic image features according to the attention mechanism to obtain the attention-weighted image features corresponding to the attention mechanism. The attention-weighted image features and the basic image features are input into the multi-head linear self-attention sub-module to obtain the attention-enhanced image, and then transmitted to the management and judgment module; It should be noted that the basic feature extraction sub-module in this embodiment includes a first convolutional layer, a second convolutional layer, and a third convolutional layer. The first convolutional layer includes two identical convolutional branches, and each convolutional branch includes 2 5X5 convolutional blocks. The first convolutional layer obtains global basic features. The second convolutional layer includes a bottleneck structure branch, an inverted bottleneck structure branch, and a residual branch. The bottleneck structure branch includes two 1X1 convolutional blocks and one 3X3 convolutional block. The inverted bottleneck structure branch includes one 1X1 convolutional block and two 3X3 convolutional blocks. The second convolutional layer obtains depth basic features. The third convolutional layer includes a multi-scale convolutional branch and a residual branch. The multi-scale convolutional branch includes one 1X1 convolutional block, one 2X2 convolutional block, and one 3X3 convolutional block. The third convolutional layer obtains multi-scale basic features. The global basic features, the depth basic features, and the multi-scale basic features are fused to obtain the basic image features.

[0024] It should be noted that in the spatial attention mechanism path of the dual-channel attention weighting sub-module in this embodiment, spatial attention weighting processing is performed on the basic image features according to the spatial attention mechanism. The spatial attention mechanism path of the dual-channel attention weighting sub-module includes a global average pooling layer, a global maximum pooling layer, one 3X3 convolutional block, one 5X5 convolutional block, and one 7X7 convolutional block. The formula for the spatial attention weighting processing is as follows: Among them, represents the basic image features, represents the spatial attention-weighted image features, represents the activation function, represents the convolutional block, represents the concatenation, represents the global maximum pooling layer, Denote the global average pooling layer to obtain the spatially attention-weighted image features. In the channel attention mechanism path of the dual-path attention-weighted sub-module in this embodiment, the channel attention-weighted processing of the basic image features is performed according to the channel attention mechanism. The channel attention mechanism path of the dual-path attention-weighted sub-module includes a global average pooling layer, a global max pooling layer, and a multi-layer perceptron. The formula for the channel attention-weighted processing is as follows: Wherein, Denote the basic image features, Denote the channel attention-weighted image features, Denote the activation function, Denote the multi-layer perceptron, Denote the global max pooling layer, Denote the global average pooling layer to obtain the channel attention-weighted image features.

[0025] It should be noted that in the linear self-attention sub-module of this embodiment, the spatially attention-weighted image features, the channel attention-weighted image features, and the basic image features are input into the linear self-attention sub-module and are respectively used as the Q matrix, K matrix, and V matrix of the linear self-attention sub-module to obtain the attention-enhanced image features. The mechanism of the linear self-attention sub-module is as follows: Wherein, Denote Q the matrix weight, Denote K the matrix weight, Denote V the matrix weight, Denote the depth convolution, Denote Q the dimension of the matrix and K the matrix, Denote the transposed matrix.

[0026] Step S03: The management judgment module judges the construction standard of the attention-enhanced image according to the preset construction database. The preset construction database includes construction drawing data, construction personnel data, and construction safety standard data. If it is determined that the current construction status does not meet the construction standard, a construction problem alarm is issued.

[0027] It should be noted that after the management judgment module in this embodiment obtains the attention-enhanced image features, it judges whether the current construction of the ultra-high-risk and major project complies with the construction standards according to the data labels in the preset construction database. The construction data labels include the construction area label and construction structure label in the construction drawing data, the construction personnel quantity label and construction personnel equipment label in the construction personnel data, and the construction dangerous area label and construction personnel safety equipment label in the construction safety standard data. The construction area label, construction structure label, construction personnel quantity label, and construction personnel equipment label are safety labels, and the construction dangerous area label and construction personnel safety equipment label are danger labels. If the safety label is missing or the danger label exists in the attention-enhanced image, an alarm for problems in the construction of the ultra-high-risk and major project will be issued.

[0028] In summary, according to the above-mentioned automated management method for extremely dangerous and major projects, by first preprocessing the monitoring images at the construction site to adapt to the complex construction environment of extremely dangerous and major projects and ensuring accurate monitoring in case of insufficient light or smoke and haze, and then designing a dual-path scenario awareness network module, including a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module, the quality of the basic image features is improved to enhance the overall monitoring accuracy. At the same time, the wide-area relationship and context relationship between image features are guided and enhanced, realizing accurate monitoring of complex multi-targets in a wide area. A management judgment module is also designed to achieve the adjustability of monitoring, avoiding false alarms or misreports caused by different construction standards and construction conditions between various extremely dangerous and major projects, and further improving the monitoring accuracy. The present invention realizes the automated monitoring management of extremely dangerous and major projects, greatly improving the monitoring accuracy in the case of a wide-area monitoring range.Specifically, the acquisition module obtains the construction site data in real time. The construction site data includes construction site image data, construction site brightness data, and construction site haze data. The construction site image data is preprocessed and then transmitted to the dual-path scenario awareness network module, eliminating the influence of dim environments and smoky haze environments, and avoiding the decline in monitoring accuracy caused by the complex construction environment of extremely dangerous and large projects. The dual-path scenario awareness network module includes a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module. The dual-path attention weighting sub-module includes a spatial attention mechanism path and a channel attention mechanism path. The basic feature extraction sub-module obtains the basic image features of the construction site image data, and performs targeted extraction of basic features from the aspects of global wide-area scale, multi-local detail scale, and deep scale, improving the quality of the basic features, and thus improving the overall monitoring accuracy. The dual-path attention weighting sub-module performs attention weighting processing on the basic image features according to the attention mechanism to obtain the attention-weighted image features corresponding to the attention mechanism, enhancing the feature quality of a single target person and target area in the case of multiple targets in a large range, and thus improving the monitoring accuracy of a single target. The attention-weighted image features and the basic image features are input into the multi-head linear self-attention sub-module to obtain the attention-enhanced image, and then transmitted to the management judgment module, enhancing the feature relationship of multiple target persons and target areas in the wide-area range, and thus improving the monitoring accuracy in the wide-area range. The management judgment module makes a construction standard judgment on the attention-enhanced image according to the preset construction database. The preset construction database includes construction drawing data, construction personnel data, and construction safety standard data. If it is determined that the current construction state does not conform to the construction standard, a construction problem alarm is issued, realizing the adjustability of monitoring and avoiding false alarms or misreports caused by different construction standards and construction conditions among multiple extremely dangerous and large projects, further improving the monitoring accuracy. The present invention realizes the automatic monitoring management of extremely dangerous and large projects, greatly improving the monitoring accuracy in the case of a wide-area monitoring range.

[0029] Please refer to Figure 2 , which shows the flowchart of the automatic monitoring management method for extremely dangerous and large projects proposed in the second embodiment of the present invention. This automatic monitoring management method for extremely dangerous and large projects includes steps S11 to S16, where: Step S11: The acquisition module obtains the construction site data in real time and performs image enhancement preprocessing. The construction site data includes construction site image data, construction site light intensity data, and construction site haze degree data. It is judged whether the construction site light intensity is less than the preset light intensity threshold according to the construction site light intensity data. If it is determined that the construction site light intensity is less than the preset light intensity threshold, light intensity enhancement preprocessing is performed on the construction site image data. It is judged whether the construction site haze degree is greater than the preset haze degree threshold according to the construction site haze degree data. If it is determined that the construction site haze degree is greater than the preset haze degree threshold, haze removal preprocessing is performed on the construction site image data; It should be noted that the preset light intensity threshold in this embodiment adopts the construction environment light visibility requirements of the ultra-dangerous and major project emergency plan, and the preset haze degree threshold adopts the construction environment soot and haze requirements of the ultra-dangerous and major project dust standard and the ultra-dangerous and major project weather emergency plan.

[0030] Step S12: Input the preprocessed construction site image data into the basic feature extraction sub-module. The basic feature extraction sub-module includes a first convolutional layer, a second convolutional layer, and a third convolutional layer. The first convolutional layer obtains global basic features, the second convolutional layer obtains depth basic features, and the third convolutional layer obtains multi-scale basic features. The global basic features, depth basic features, and multi-scale basic features are fused to obtain basic image features; Step S13: Input the basic image features into the spatial attention mechanism path of the dual-path attention weighting sub-module, and perform spatial attention weighting processing on the basic image features according to the spatial attention mechanism to obtain spatially attention-weighted image features; Step S14: Input the basic image features into the channel attention mechanism path of the dual-path attention weighting sub-module, and perform channel attention weighting processing on the basic image features according to the channel attention mechanism to obtain channel attention-weighted image features; Step S15: Input the spatially attention-weighted image features, channel attention-weighted image features, and basic image features into the linear self-attention sub-module, and use them as the Q matrix, K matrix, and V matrix of the linear self-attention sub-module respectively to obtain attention-enhanced image features, and transmit the obtained attention-enhanced image to the management and judgment module; Step S16: After the management judgment module obtains the attention-enhanced image features, it determines whether the current construction of the ultra-high risk and major project complies with the construction standards according to the data labels in the preset construction database. The construction area label, construction structure label, number of construction personnel label, and construction personnel's used equipment label are safety labels, and the construction dangerous area label and construction personnel's safety equipment label are danger labels. If a safety label is missing or a danger label exists in the attention-enhanced image, an alarm for problems in the construction of the ultra-high risk and major project is issued.

[0031] In summary, according to the above-mentioned automated management method for extremely dangerous and major projects, by first preprocessing the monitoring images at the construction site to adapt to the complex construction environment of extremely dangerous and major projects and ensure accurate monitoring in case of insufficient light or dust and haze, and then designing a dual-path scenario awareness network module, including a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module, the quality of the basic image features is improved to enhance the overall monitoring accuracy. At the same time, the wide-area relationship and context relationship between image features are guided and enhanced, realizing accurate monitoring of complex multi-targets in a wide area. A management judgment module is also designed to achieve the adjustability of monitoring, avoiding false alarms or misreports caused by different construction standards and construction conditions among various extremely dangerous and major projects, and further improving the monitoring accuracy. The present invention realizes the automated monitoring management of extremely dangerous and major projects, greatly improving the monitoring accuracy in the case of a wide-area monitoring range.Specifically, the acquisition module obtains the construction site data in real time. The construction site data includes construction site image data, construction site luminance data, and construction site haze data. The construction site image data is preprocessed and then transmitted to the dual-path scenario perception network module, eliminating the influence of dim environments and smoky and hazy environments, and avoiding the decline in monitoring accuracy caused by the complex construction environment of extremely dangerous and large-scale projects. The dual-path scenario perception network module includes a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module. The dual-path attention weighting sub-module includes a spatial attention mechanism path and a channel attention mechanism path. The basic feature extraction sub-module obtains the basic image features of the construction site image data, and conducts targeted extraction of basic features from the aspects of global wide-area scale, multi-local detail scale, and deep scale, improving the quality of the basic features, and thus improving the overall monitoring accuracy. The dual-path attention weighting sub-module performs attention weighting processing on the basic image features according to the attention mechanism to obtain the attention-weighted image features corresponding to the attention mechanism, enhancing the feature quality of a single target person and target area in the case of a large range of multiple targets, and thus improving the monitoring accuracy of a single target. The attention-weighted image features and the basic image features are input into the multi-head linear self-attention sub-module to obtain an attention-enhanced image, which is then transmitted to the management judgment module, enhancing the feature relationship of multiple target persons and target areas in the wide-area range, and thus improving the monitoring accuracy in the wide-area range. The management judgment module makes a construction standard judgment on the attention-enhanced image according to the preset construction database. The preset construction database includes construction drawing data, construction personnel data, and construction safety standard data. If it is determined that the current construction state does not conform to the construction standard, a construction problem alarm is issued, realizing the adjustability of monitoring and avoiding false alarms or misreports caused by different construction standards and construction conditions among multiple extremely dangerous and large-scale projects, further improving the monitoring accuracy. The present invention realizes the automated monitoring management of extremely dangerous and large-scale projects, greatly improving the monitoring accuracy in the case of a wide-area monitoring range.

[0032] Please refer to Figure 3 , which shows a schematic structural diagram of the automated monitoring management system for extremely dangerous and large-scale projects proposed in the third embodiment of the present invention. The system includes: An acquisition module 10, configured to acquire the construction site data in real time. The construction site data includes construction site image data, construction site luminance data, and construction site haze data. The construction site image data is preprocessed and then transmitted to the dual-path scenario perception network module; The dual-path scenario awareness network module 20, where the dual-path scenario awareness network module includes a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module. The dual-path attention weighting sub-module includes a spatial attention mechanism path and a channel attention mechanism path. The basic feature extraction sub-module obtains the basic image features of the construction site image data. The dual-path attention weighting sub-module performs attention weighting processing on the basic image features according to the attention mechanism to obtain the attention-weighted image features corresponding to the attention mechanism, and inputs the attention-weighted image features and the basic image features into the multi-head linear self-attention sub-module to obtain the attention-enhanced image, and transmits it to the management judgment module; The management judgment module 30, where the management judgment module performs construction standard judgment on the attention-enhanced image according to the preset construction database. The preset construction database includes construction drawing data, construction personnel data, and construction safety standard data. If it is determined that the current construction state does not meet the construction standards, a construction problem alarm is issued.

[0033] Furthermore, the acquisition module 10 includes: The acquisition preprocessing unit 101, which is used for the acquisition module to obtain the construction site data in real time. The construction site data includes construction site image data, construction site luminance data, and construction site haze data, preprocesses the construction site image data, and then transmits it to the dual-path scenario awareness network module.

[0034] Furthermore, the dual-path scenario awareness network module 20 includes: The basic feature extraction unit 201, where the dual-path scenario awareness network module includes a basic feature extraction sub-module, a dual-channel attention weighting sub-module, and a multi-head linear self-attention sub-module. The dual-path attention weighting sub-module includes a spatial attention mechanism path and a channel attention mechanism path. The basic feature extraction sub-module obtains the basic image features of the construction site image data; The dual-path attention weighting unit 202, which is used for the dual-path attention weighting sub-module to perform attention weighting processing on the basic image features according to the attention mechanism to obtain the attention-weighted image features corresponding to the attention mechanism; The multi-head linear self-attention unit 203, which is used to input the attention-weighted image features and the basic image features into the multi-head linear self-attention sub-module to obtain the attention-enhanced image, and transmits it to the management judgment module.

[0035] Furthermore, the management judgment module 30 includes: The management judgment unit 301 is configured to enable the management judgment module to judge the construction standard of the attention-enhanced image according to a preset construction database, where the preset construction database includes construction drawing data, construction personnel data, and construction safety standard data. If it is determined that the current construction state does not meet the construction standard, a construction problem alarm is issued.

[0036] The present invention also provides a computer storage medium, on which one or more programs are stored, and when the program is executed by a processor, the above-mentioned ultra-dangerous and major project automatic monitoring and management method is implemented.

[0037] The present invention also provides a computer device, including a memory and a processor, where the memory is used to store a computer program, and the processor is used to execute the computer program stored on the memory to implement the above-mentioned ultra-dangerous and major project automatic monitoring and management method.

[0038] Those skilled in the art can understand that the logic and / or steps represented in the flowchart or described in other ways herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in combination with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

[0039] More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection portion with one or more wirings (electronic device), a portable computer disk cartridge (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or other appropriate processing as necessary, and then stored in a computer memory.

[0040] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), and the like.

[0041] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0042] The above-described embodiments merely represent several implementation manners of the present invention. The description thereof is relatively specific and detailed, but should not be construed as a limitation on the scope of the patent of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention should be subject to the appended claims.

Claims

1. An automated management method for ultra-hazardous large projects, characterized in that: include: The acquisition module acquires the construction site data in real time, the construction site data including the construction site image data, the construction site brightness data and the construction site haze data, pre-processes the construction site image data, and then transmits it to the dual-path scenario perception network module; The dual-path scenario perception network module includes a basic feature extraction submodule, a dual-channel attention weighting submodule and a multi-head linear self-attention submodule. The dual-path attention weighting submodule includes a spatial attention mechanism path and a channel attention mechanism path. The basic feature extraction submodule obtains the basic image features of the construction site image data. The dual-path attention weighting submodule performs attention weighting processing on the basic image features according to the attention mechanism to obtain attention weighted image features corresponding to the attention mechanism. The attention weighted image features and the basic image features are input into the multi-head linear self-attention submodule to obtain an attention-enhanced image, and the image is transmitted to the management judgment module. The management judgment module performs construction standard judgment on the attention-enhanced image based on a preset construction database, which includes construction drawing data, construction personnel data and construction safety standard data. If it is determined that the current construction status does not meet the construction standards, a construction problem alarm is issued.

2. The method for automated management of ultra-hazardous and large projects according to claim 1 is characterized in that: The acquisition module acquires construction site data in real time, the construction site data including construction site image data, construction site brightness data and construction site haze data, and the step of preprocessing the construction site image data includes: The acquisition module acquires the construction site data in real time and performs image enhancement preprocessing, wherein the construction site data includes construction site image data, construction site brightness data and construction site haze data; According to the construction site brightness data, it is determined whether the brightness of the construction site is less than a preset brightness threshold. If it is determined that the brightness of the construction site is less than the preset brightness threshold, the construction site image data is pre-processed for brightness enhancement, pixel values ​​of the red, green and blue channels in the construction site image data are extracted, and weighted grayscale processing is performed on the construction site image data according to the following formula: in, represents the weighted gray value of each pixel, a represents the total number of pixels, Represents the maximum value of the three-channel pixel value of each pixel The grayscale weight value, Represents the minimum value of the three-channel pixel values ​​of each pixel The grayscale weight value of R, G, and B respectively represents the pixel value of the red channel, green channel, and blue channel of each pixel point. After the construction site image data is linearly normalized, the image grayscale mean value processing is performed according to the following formula: in, Representing the grayscale mean of the pixel points, and taking the grayscale mean as the image grayscale value of the construction site image data; According to the construction site haze data, it is determined whether the construction site haze is greater than a preset haze threshold. If it is determined that the construction site haze is greater than the preset haze threshold, the construction site image data is pre-processed for haze removal, and the atmospheric light value A is estimated according to the image gray value of the construction site image data. The estimated atmospheric light value A is calculated and guided by filtering to obtain the transmittance. , according to the atmospheric light value A and the transmittance A preprocessed image is obtained, and the formula for obtaining the preprocessed image is as follows: in, represents the preprocessed image, represents an unprocessed image, represents the lower limit of transmittance, and max(·) represents obtaining the maximum value.

3. The method for automated management of ultra-hazardous and large projects according to claim 1 is characterized in that: The step of the basic feature extraction submodule acquiring the basic image features of the construction site image data comprises: The preprocessed construction site image data is input into a basic feature extraction submodule, which includes a first convolution layer, a second convolution layer and a third convolution layer. The first convolution layer includes two identical convolution branches, each of which includes two 5X5 convolution blocks. The first convolution layer obtains global basic features. The second convolution layer includes a bottleneck structure branch, an inverse bottleneck structure branch and a residual branch. The bottleneck structure branch includes two 1X1 convolution blocks and one 3X3 convolution block. The inverse bottleneck structure branch includes a 1X1 convolution block and two 3X3 convolution blocks. The second convolution layer obtains deep basic features. The third convolution layer includes a multi-scale convolution branch and a residual branch. The multi-scale convolution branch includes a 1X1 convolution block, a 2X2 convolution block and a 3X3 convolution block. The third convolution layer obtains multi-scale basic features. The global basic features, the deep basic features and the multi-scale basic features are fused to obtain basic image features.

4. The method for automated management of ultra-hazardous and large projects according to claim 1 is characterized in that: The dual-path attention weighted submodule performs attention weighting processing on the basic image features according to the attention mechanism to obtain the attention weighted image features corresponding to the attention mechanism, including: The basic image features are input into the spatial attention mechanism path of the dual-path attention weighted submodule, and the basic image features are subjected to spatial attention weighted processing according to the spatial attention mechanism. The spatial attention mechanism path of the dual-path attention weighted submodule includes a global average pooling layer, a global maximum pooling layer, a 3X3 convolution block, a 5X5 convolution block and a 7X7 convolution block. The formula for the spatial attention weighted processing is as follows: in, represents the basic image features, represents the spatial attention weighted image features, represents the activation function, represents the convolutional block, Indicates splicing, represents the global maximum pooling layer, represents the global average pooling layer; To obtain spatial attention weighted image features.

5. The method for automated management of ultra-hazardous and large projects according to claim 1 is characterized in that: The dual-path attention weighted submodule performs attention weighting processing on the basic image features according to the attention mechanism to obtain the attention weighted image features corresponding to the attention mechanism, and the step also includes: The basic image features are input into the channel attention mechanism path of the dual-path attention weighted submodule, and the basic image features are subjected to channel attention weighted processing according to the channel attention mechanism. The channel attention mechanism path of the dual-path attention weighted submodule includes a global average pooling layer, a global maximum pooling layer and a multi-layer perceptron. The channel attention weighted processing is shown as follows: in, represents the basic image features, represents the channel attention weighted image features, represents the activation function, represents a multilayer perceptron, represents the global maximum pooling layer, represents the global average pooling layer; To obtain channel attention weighted image features.

6. The method for automated management of ultra-hazardous and large projects according to claim 1 is characterized in that: The step of inputting the attention-weighted image features and the basic image features into the multi-head linear self-attention submodule to obtain an attention-enhanced image, and transmitting the image to the management judgment module includes: The spatial attention weighted image features, channel attention weighted image features and basic image features are input into the linear self-attention submodule and used as the linear self-attention submodule respectively. Q matrix, K Matrix and V Matrix to obtain attention-enhanced image features. The mechanism of the linear self-attention submodule is as follows: in, express Q Matrix weights, express K Matrix weights, express V Matrix weights, represents the depthwise convolution, express Q Matrix and K The dimension of the matrix, represents the transposed matrix; The acquired attention-enhanced image is transmitted to the management judgment module.

7. The method for automated management of ultra-hazardous and large projects according to claim 1 is characterized in that: The step of the management judgment module judging the construction standard of the attention-enhanced image according to the preset construction database includes: After the management judgment module obtains the attention-enhanced image features, it judges whether the current super-dangerous and large-scale construction project meets the construction standards based on the data labels in the preset construction database. The construction data labels include the construction area label and the construction structure label in the construction drawing data, and the number of construction personnel label and the equipment used by construction personnel label in the construction personnel data, as well as the construction danger area label and the construction personnel safety equipment label in the construction safety standard data. The construction area label, the construction structure label, the number of construction personnel label and the equipment used by construction personnel label are safety labels, and the construction danger area label and the construction personnel safety equipment label are danger labels. If the safety label is missing or a danger label exists in the attention-enhanced image, an alarm for super-dangerous and large-scale construction problems will be issued.

8. An automated management system for ultra-hazardous and large projects, characterized in that: include: An acquisition module is used to acquire construction site data in real time, wherein the construction site data includes construction site image data, construction site brightness data and construction site haze data, pre-process the construction site image data, and then transmit it to the dual-path scenario perception network module; A dual-path scenario-aware network module, wherein the dual-path scenario-aware network module includes a basic feature extraction submodule, a dual-channel attention weighting submodule and a multi-head linear self-attention submodule, wherein the dual-path attention weighting submodule includes a spatial attention mechanism path and a channel attention mechanism path, wherein the basic feature extraction submodule obtains basic image features of the construction site image data, wherein the dual-path attention weighting submodule performs attention weighting processing on the basic image features according to the attention mechanism to obtain attention weighted image features corresponding to the attention mechanism, wherein the attention weighted image features and the basic image features are input into the multi-head linear self-attention submodule to obtain an attention-enhanced image, and transmit the image to the management judgment module; A management judgment module is used for judging the construction standards of the attention-enhanced image according to a preset construction database, wherein the preset construction database includes construction drawing data, construction personnel data and construction safety standard data. If it is determined that the current construction status does not meet the construction standards, a construction problem alarm is issued.

9. A storage medium, characterized in that: The storage medium stores one or more programs, which, when executed by the processor, implement the method for automated management of ultra-hazardous and large projects as described in any one of claims 1 to 7.

10. A computer device, characterized in that: The computer device comprises a memory and a processor, wherein: The memory is used to store computer programs; When the processor is used to execute the computer program stored in the memory, the method for automated management of ultra-hazardous and large projects described in any one of claims 1 to 7 is implemented.

Images