CN117061711B - Video monitoring safety management method and system based on Internet of things - Google Patents
Video monitoring safety management method and system based on Internet of thingsInfo
- Publication number
- CN117061711B CN117061711B CN202311308817.0A CN202311308817A CN117061711B CN 117061711 B CN117061711 B CN 117061711B CN 202311308817 A CN202311308817 A CN 202311308817A CN 117061711 B CN117061711 B CN 117061711B
- Authority
- CN
- China
- Prior art keywords
- data
- video monitoring
- monitoring
- video
- module
- 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.)
- Active
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 267
- 238000007726 management method Methods 0.000 title claims abstract description 21
- 238000012545 processing Methods 0.000 claims abstract description 40
- 238000004458 analytical method Methods 0.000 claims abstract description 31
- 238000007781 pre-processing Methods 0.000 claims abstract description 19
- 230000005540 biological transmission Effects 0.000 claims abstract description 15
- 238000013138 pruning Methods 0.000 claims description 51
- 238000012549 training Methods 0.000 claims description 25
- 238000000034 method Methods 0.000 claims description 22
- 238000010586 diagram Methods 0.000 claims description 16
- 239000011159 matrix material Substances 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 11
- 238000013507 mapping Methods 0.000 claims description 9
- 238000005516 engineering process Methods 0.000 claims description 8
- 230000009467 reduction Effects 0.000 claims description 8
- 230000011218 segmentation Effects 0.000 claims description 8
- 238000000605 extraction Methods 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 5
- 125000004122 cyclic group Chemical group 0.000 claims description 4
- 230000004927 fusion Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000000903 blocking effect Effects 0.000 claims description 2
- 230000008901 benefit Effects 0.000 description 4
- 238000012806 monitoring device Methods 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
Abstract
The invention relates to the technical field of the Internet of things and discloses a video monitoring safety management method and system based on the Internet of things, wherein monitoring content is acquired through video monitoring equipment to obtain first video monitoring data, and the first video monitoring data is stored in an IP-SAN; reading first video monitoring data in the IP-SAN, transmitting the first video monitoring data to an Internet of things platform, and carrying out data preprocessing on the first video monitoring data to obtain second video monitoring data; determining the application scene of the current video monitoring, constructing a corresponding monitoring processing model, inputting second video monitoring data into the monitoring processing model, and outputting a monitoring analysis result through the monitoring processing model; integrating the monitoring analysis results to generate target video monitoring content and corresponding safety early warning information, and encrypting and transmitting the target video monitoring content to a user terminal based on the safety early warning information; the invention realizes the efficient and safe transmission of video data and improves the working efficiency of video monitoring.
Description
Technical Field
The invention relates to the technical field of the Internet of things, in particular to a video monitoring safety management method and system based on the Internet of things.
Background
Along with the development of science and technology, video monitoring plays an important role, so that the safety of living environment is greatly improved, but the current video monitoring safety and transmission efficiency are lower, and therefore, the video monitoring safety management method and system based on the Internet of things are provided.
Disclosure of Invention
The invention aims to solve the problems and designs a video monitoring safety management method and equipment based on the Internet of things.
The invention provides a video monitoring safety management method based on the Internet of things, which comprises the following steps:
monitoring content acquisition is carried out through video monitoring equipment to obtain first video monitoring data, and the first video monitoring data are stored into an IP-SAN;
reading the first video monitoring data in the IP-SAN, transmitting the first video monitoring data to an Internet of things platform, and carrying out data preprocessing on the first video monitoring data to obtain second video monitoring data;
Determining an application scene of current video monitoring, constructing a corresponding monitoring processing model, inputting the second video monitoring data into the monitoring processing model, and outputting a monitoring analysis result through the monitoring processing model;
and integrating the monitoring analysis results to generate target video monitoring content and corresponding safety early warning information, and encrypting and transmitting the target video monitoring content to a user terminal based on the safety early warning information.
Optionally, in a first implementation manner of the first aspect of the present invention, performing data preprocessing on the first video monitoring data to obtain second video monitoring data, including:
Acquiring image data in the first video monitoring data, and determining that two 3×3 matrix operators are respectively used for the image data in the transverse and longitudinal directions of the first video monitoring data;
performing first-order convolution operation on matrix operators in the longitudinal and transverse directions and pixel points on the image to obtain a gradient derivative in the longitudinal and transverse directions;
Adding the first step derivative in the longitudinal and transverse directions and solving the root mean square to obtain gradient data in the longitudinal and transverse directions;
and combining the gradient data in the transverse and longitudinal directions, comparing the gradient data with a set threshold value, determining whether the data of each pixel position is completely black or completely white, and outputting first image data.
Optionally, in a second implementation manner of the first aspect of the present invention, after the combining the gradient data in the transverse and longitudinal directions, comparing the gradient data with a set threshold value, determining that the data at each pixel position is full black or full white, and outputting the first image data, the method further includes:
Acquiring the first image data, and extracting sub-bands corresponding to each frequency band of the first image data to obtain a high-frequency band sub-band and a low-frequency band sub-band;
Blocking the high-frequency sub-band and the low-frequency sub-band, and calculating image noise data based on a partial differential equation;
carrying out dimension reduction processing on the first image data according to the image noise data, and connecting noise singular points of all the block areas in the image noise data to form a linear structure;
obtaining reconstruction data of the first image data after dimension reduction to obtain the first image data, and carrying out image enhancement on the first image data by adopting wavelet transformation and a mathematical morphology method to obtain second image data;
And integrating the second image data to obtain second video monitoring data.
Optionally, in a third implementation manner of the first aspect of the present invention, the determining an application scenario of current video monitoring, and constructing a corresponding monitoring processing model includes:
acquiring a history monitoring image of a video monitoring application scene, forming a training set, and inputting the training set into YOLOv network structures;
adopting a step-by-step channel convolution mode, in a two-dimensional plane, the convolution kernels are in one-to-one correspondence with the number of channels, one convolution kernel processes one channel, and 3 feature images can be generated after the history monitoring images in the training set are processed;
the number of the feature mappings subjected to the depth convolution is consistent with the channel value of the input layer, and the feature combination is performed by adopting point-by-point convolution to obtain new feature mappings;
and (5) performing model adjustment by adopting sparse training and pruning technology to obtain a monitoring processing model.
Optionally, in a fourth implementation manner of the first aspect of the present invention, the performing model adjustment by using a sparse training and pruning technology to obtain a monitoring processing model includes:
pruning is carried out on each regularization module as a whole, pruning is carried out according to the bottom-to-high order of the weight average value of the regularization module, and pruning is carried out according to the half ratio of the regularization module in the primary pruning process;
compensating by combining with the fine adjustment of the model, and continuously carrying out cyclic iteration to obtain a network structure after layer pruning;
Determining a channel pruning threshold corresponding to the global value reduction rate according to the pruning parameters and the scaling sparseness of the regularization module, and determining a protection threshold of each neural network layer according to the channel protection threshold parameters;
and when the channel protection threshold value is larger than the pruning threshold value and smaller than the channel protection threshold value, channel pruning is carried out on the network structure after layer pruning.
Optionally, in a fifth implementation manner of the first aspect of the present invention, integrating the monitoring analysis result to generate the target video monitoring content and the corresponding security pre-warning information includes:
Acquiring a video stream of the second video monitoring data, acquiring a video frame image from the video stream, and correcting the video frame image by adopting a division model;
Normalizing the corrected video frame image, taking the normalized video frame image as the input of Deeplabv & lt3+ & gt model, extracting the characteristics layer by layer through a coding network, and obtaining the characteristic codes of the pictures;
Upsampling the feature codes through a decoding network to obtain a segmentation mask diagram with the same resolution as the image input by the model;
And video stitching fusion is carried out based on the segmentation mask diagram to obtain target video monitoring content, and safety early warning information is obtained according to the monitoring analysis result.
Optionally, in a sixth implementation manner of the first aspect of the present invention, the encrypting the target video surveillance content based on the security pre-warning information is transmitted to a user terminal, including:
receiving the target video monitoring, and extracting a header field and a content field of an original video data packet of the target video monitoring;
performing hash operation on the header field and the content field respectively to obtain a corresponding hash value field;
Forming a new video data packet by the hash value field corresponding to the original video data packet and the content field of the original video data packet;
and obtaining a new data stream of the target video monitoring according to the new video data packet, and encrypting the new data stream of the target video monitoring by adopting a cross chaos algorithm.
The invention provides a steep slope surface gradient detection device based on image processing, wherein the video monitoring safety management system based on the Internet of things comprises a content acquisition module, a data preprocessing module, a monitoring analysis module and an encryption transmission module, wherein the content acquisition module is used for acquiring monitoring content through the video monitoring device to obtain first video monitoring data, and storing the first video monitoring data into an IP-SAN;
The data preprocessing module is used for reading the first video monitoring data in the IP-SAN, transmitting the first video monitoring data to an Internet of things platform, and preprocessing the first video monitoring data to obtain second video monitoring data;
the monitoring analysis module is used for determining the application scene of the current video monitoring, constructing a corresponding monitoring processing model, inputting the second video monitoring data into the monitoring processing model, and outputting a monitoring analysis result through the monitoring processing model;
And the encryption transmission module is used for integrating the monitoring analysis results, generating target video monitoring content and corresponding safety early warning information, and encrypting and transmitting the target video monitoring content to a user terminal based on the safety early warning information.
Optionally, in a sixth implementation manner of the first aspect of the present invention, the content acquisition module includes an acquisition sub-module, a convolution operation sub-module, a calculation sub-module, and a data combination sub-module, where the acquisition sub-module is configured to acquire image data in the first video monitoring data, and determine that two 3×3 matrix operators are respectively used for image data in a transverse and longitudinal direction of the first video monitoring data;
The convolution operation sub-module is used for carrying out first-order convolution operation on the matrix operator in the longitudinal and transverse directions and the pixel points on the image to obtain a gradient derivative in the longitudinal and transverse directions;
The computing sub-module is used for adding the first step derivative in the longitudinal and transverse directions and solving the root mean square to obtain gradient data in the longitudinal and transverse directions;
And the data combining sub-module is used for combining the gradient data in the transverse and longitudinal directions, comparing the gradient data with a set threshold value, determining whether the data of each pixel position is completely black or completely white, and outputting first image data.
Optionally, in a sixth implementation manner of the first aspect of the present invention, the encryption transmission module includes an extraction sub-module, a hash operation sub-module, a composition sub-module, and an encryption sub-module, where the extraction sub-module is configured to receive the target video surveillance and extract a header field and a content field of an original video data packet of the target video surveillance;
The hash operation sub-module is used for carrying out hash operation on the header field and the content field respectively to obtain a corresponding hash value field;
the composition sub-module is used for composing the hash value field corresponding to the original video data packet and the content field of the original video data packet into a new video data packet;
And the encryption sub-module is used for obtaining the new data stream of the target video monitoring according to the new video data packet, and encrypting the new data stream of the target video monitoring by adopting a cross chaos algorithm.
In the technical scheme provided by the invention, monitoring content is acquired through video monitoring equipment to obtain first video monitoring data, and the first video monitoring data is stored in an IP-SAN; reading the first video monitoring data in the IP-SAN, transmitting the first video monitoring data to an Internet of things platform, and carrying out data preprocessing on the first video monitoring data to obtain second video monitoring data; determining an application scene of current video monitoring, constructing a corresponding monitoring processing model, inputting the second video monitoring data into the monitoring processing model, and outputting a monitoring analysis result through the monitoring processing model; integrating the monitoring analysis results to generate target video monitoring content and corresponding safety early warning information, and encrypting and transmitting the target video monitoring content to a user terminal based on the safety early warning information; the invention realizes the efficient and safe transmission of video data and improves the working efficiency of video monitoring.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention.
Fig. 1 is a schematic diagram of a first embodiment of a video monitoring security management method based on internet of things according to an embodiment of the present invention;
Fig. 2 is a schematic diagram of a second embodiment of a video monitoring security management method based on internet of things according to an embodiment of the present invention;
Fig. 3 is a schematic diagram of a third embodiment of a video monitoring security management method based on internet of things according to an embodiment of the present invention;
Fig. 4 is a schematic structural diagram of a video monitoring security management system based on the internet of things according to an embodiment of the present invention;
fig. 5 is a schematic structural diagram of another video monitoring security management system based on the internet of things according to an embodiment of the present invention.
Detailed Description
The terms "first," "second," "third," "fourth" and the like in the description and in the claims and in the above drawings, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments described herein may be implemented in other sequences than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, apparatus, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus.
For convenience of understanding, a specific flow of the embodiment of the present invention is described below, referring to fig. 1, which is a schematic diagram of a first embodiment of a video monitoring security management method based on the internet of things, where a video monitoring device is connected to an internet of things platform through the internet of things, and the method includes the following steps:
Step 101, monitoring content acquisition is carried out through video monitoring equipment to obtain first video monitoring data, and the first video monitoring data are stored in an IP-SAN;
In this embodiment, the IP-SAN refers to Internet Protocol Storage Area Network, i.e., a storage area network based on the IP protocol. The IP network is used for transmitting the storage data, so that the data interaction between the storage device and the server is realized. The main advantage of the IP-SAN is that it utilizes existing IP network facilities, reducing construction and maintenance costs, while also providing flexibility and scalability.
102, Reading first video monitoring data in an IP-SAN, transmitting the first video monitoring data to an Internet of things platform, and carrying out data preprocessing on the first video monitoring data to obtain second video monitoring data;
Step 103, determining the application scene of the current video monitoring, constructing a corresponding monitoring processing model, inputting second video monitoring data into the monitoring processing model, and outputting a monitoring analysis result through the monitoring processing model;
In the embodiment, a history monitoring image of a video monitoring application scene is obtained, a training set is formed, and the training set is input into YOLOv network structure; adopting a step-by-step channel convolution mode, in a two-dimensional plane, the convolution kernels are in one-to-one correspondence with the number of channels, one convolution kernel processes one channel, and 3 feature images can be generated after the history monitoring images in the training set are processed; the number of the feature mappings subjected to the depth convolution is consistent with the channel value of the input layer, and the feature combination is performed by adopting point-by-point convolution to obtain new feature mappings; and (5) performing model adjustment by adopting sparse training and pruning technology to obtain a monitoring processing model.
In the embodiment, each regularization module is used as a whole for pruning, pruning is performed according to the bottom-to-high order of the weight average value of the regularization module, and pruning is performed according to the half ratio of the regularization module in the primary pruning process; compensating by combining with the fine adjustment of the model, and continuously carrying out cyclic iteration to obtain a network structure after layer pruning; determining a channel pruning threshold corresponding to the global value reduction rate according to the pruning parameters and the scaling sparseness of the regularization module, and determining a protection threshold of each neural network layer according to the channel protection threshold parameters; and when the channel protection threshold value is larger than the pruning threshold value and smaller than the channel protection threshold value, channel pruning is carried out on the network structure after layer pruning.
In this embodiment, sparse training reduces model complexity: by adding a sparse loss in the training process, we force the neural network to train out sparse parameters, which means that many parameters are close to or equal to zero in the model, so this can be regarded as a feature selection method, because a smaller parameter value means that the influence of the corresponding input feature on the model is smaller, and the sparse parameters can prevent the model from being too much dependent on the input feature, so that the complexity of the model is reduced; sparse training improves model generalization capability: the generalization capability refers to the prediction performance of a model for unknown data, and a model with high generalization capability can still have good performance when new data is encountered, and the generalization capability of the model can be improved by adding sparse loss into a conventional loss function through the following ways; sparse training prevents overfitting: overfitting means that the model is so complex that it performs well on training data, but performs poorly on test data, and by sparsifying parameters we can reduce the complexity of the model, thereby reducing the risk of overfitting; sparse training improves model robustness: the sparsification parameters may make the model exhibit a higher robustness to noise and unimportant features. This is because the sparsification of the parametric forces the model to focus on a small number of key features, rather than being sensitive to each input feature, so that the predictive performance of the model is not greatly affected when new data is encountered that contains noise or unimportant features.
And 104, integrating the monitoring analysis results to generate target video monitoring content and corresponding safety early warning information, and encrypting and transmitting the target video monitoring content to the user terminal based on the safety early warning information.
In this embodiment, a video stream of second video monitoring data is obtained, a video frame image is obtained from the video stream, and a division model is used to correct the video frame image; normalizing the corrected video frame image, taking the normalized video frame image as the input of Deeplabv & lt3+ & gt model, extracting the characteristics layer by layer through a coding network, and obtaining the characteristic codes of the pictures; upsampling the feature codes through a decoding network to obtain a segmentation mask diagram with the same resolution as the image input by the model; and video stitching fusion is carried out based on the segmentation mask diagram, target video monitoring content is obtained, and safety early warning information is obtained according to a monitoring analysis result.
In the embodiment of the invention, monitoring content is acquired through video monitoring equipment to obtain first video monitoring data, and the first video monitoring data is stored in an IP-SAN; reading first video monitoring data in the IP-SAN, transmitting the first video monitoring data to an Internet of things platform, and carrying out data preprocessing on the first video monitoring data to obtain second video monitoring data; determining the application scene of the current video monitoring, constructing a corresponding monitoring processing model, inputting second video monitoring data into the monitoring processing model, and outputting a monitoring analysis result through the monitoring processing model; integrating the monitoring analysis results to generate target video monitoring content and corresponding safety early warning information, and encrypting and transmitting the target video monitoring content to a user terminal based on the safety early warning information; the invention realizes the efficient and safe transmission of video data and improves the working efficiency of video monitoring.
Referring to fig. 2, a second embodiment of a video monitoring security management method based on internet of things according to an embodiment of the present invention is shown, where the method includes:
step 201, acquiring image data in first video monitoring data, and determining that two 3×3 matrix operators are respectively used for the image data in the transverse and longitudinal directions of the first video monitoring data;
Step 202, performing first-order convolution operation on pixel points on an image by utilizing a matrix operator in the longitudinal and transverse directions to obtain a gradient derivative in the longitudinal and transverse directions;
Step 203, adding the derivative of one step in the longitudinal and transverse directions and solving the root mean square to obtain gradient data in the longitudinal and transverse directions;
and 204, combining the gradient data in the transverse and longitudinal directions, comparing the gradient data with a set threshold value, determining whether the data of each pixel position is completely black or completely white, and outputting first image data.
Step 205, acquiring first image data, and extracting sub-bands corresponding to each frequency band of the first image data to obtain a high-frequency band sub-band and a low-frequency band sub-band;
Step 206, the high-frequency sub-band and the low-frequency sub-band are segmented, and image noise data are calculated based on partial differential equations;
Step 207, carrying out dimension reduction processing on the first image data according to the image noise data, and connecting noise singular points of all the block areas in the image noise data to form a linear structure;
step 208, obtaining reconstruction data of the first image data after dimension reduction to obtain first image data, and carrying out image enhancement on the first image data by adopting wavelet transformation and a mathematical morphology method to obtain second image data;
and step 209, integrating the second image data to obtain second video monitoring data.
Referring to fig. 3, a third embodiment of a video monitoring security management method based on internet of things according to an embodiment of the present invention is shown, where the method includes:
Step 301, receiving target video monitoring, and extracting a header field and a content field of an original video data packet of the target video monitoring;
Step 302, performing hash operation on the header field and the content field respectively to obtain a corresponding hash value field;
In this embodiment, the hash is to transform an input of any length (also called as pre-mapped pre-image) into an output of a fixed length, which is a hash value, through a hash algorithm; this conversion is a compressed mapping, i.e. the hash value is typically much smaller in space than the input, different inputs may be hashed to the same output, so it is not possible to determine a unique input value from the hash value, as a function of compressing a message of arbitrary length to a message digest of a certain fixed length.
Step 303, forming a new video data packet from the hash value field corresponding to the original video data packet and the content field of the original video data packet;
And step 304, obtaining a new data stream of the target video monitoring according to the new video data packet, and encrypting the new data stream of the target video monitoring by adopting a cross chaos algorithm.
Referring to fig. 4, a schematic structural diagram of a video monitoring security management system based on internet of things provided by an embodiment of the present invention includes a content acquisition module, a data preprocessing module, a monitoring analysis module and an encryption transmission module, where the content acquisition module 401 is configured to acquire monitoring content through a video monitoring device, obtain first video monitoring data, and store the first video monitoring data into an IP-SAN;
The data preprocessing module 402 is configured to read first video monitoring data in the IP-SAN, transmit the first video monitoring data to the internet of things platform, and perform data preprocessing on the first video monitoring data to obtain second video monitoring data;
the monitoring analysis module 403 is configured to determine an application scenario of current video monitoring, construct a corresponding monitoring processing model, input second video monitoring data into the monitoring processing model, and output a monitoring analysis result through the monitoring processing model;
and the encryption transmission module 404 is used for integrating the monitoring analysis results, generating target video monitoring content and corresponding safety early warning information, and encrypting and transmitting the target video monitoring content to the user terminal based on the safety early warning information.
Referring to fig. 5, another structural schematic diagram of a video monitoring security management system based on internet of things in an embodiment of the present invention includes:
The content acquisition module 401 comprises an acquisition submodule, a convolution operation submodule, a calculation submodule and a data combination submodule, wherein the acquisition submodule 4011 is used for acquiring image data in first video monitoring data and determining that two 3×3 matrix operators are respectively used for the image data in the transverse and longitudinal directions of the first video monitoring data;
A convolution operation sub-module 4012, configured to perform a first-order convolution operation with the pixel points on the image by using a matrix operator in the longitudinal and transverse directions, so as to obtain a first-step derivative in the longitudinal and transverse directions;
a calculating submodule 4013 for adding up the first step derivative in the longitudinal and transverse directions and obtaining the root mean square to obtain the gradient data in the longitudinal and transverse directions;
the data combining submodule 4014 is configured to combine the gradient data in the transverse and longitudinal directions, compare the gradient data with a set threshold value, determine whether the data at each pixel position is full black or full white, and output first image data.
The encryption transmission module 404 comprises an extraction submodule, a hash operation submodule, a composition submodule and an encryption submodule, wherein the extraction submodule 4041 is used for receiving target video monitoring and extracting a header field and a content field of an original video data packet of the target video monitoring;
a hash operation sub-module 4042, configured to perform hash operations on the header field and the content field respectively, to obtain a corresponding hash value field;
a composing sub-module 4043, configured to compose a new video data packet from the hash value field corresponding to the original video data packet and the content field of the original video data packet;
and the encryption submodule 4044 is used for obtaining a new data stream of the target video monitoring according to the new video data packet and encrypting the new data stream of the target video monitoring by adopting a cross chaos algorithm.
Through implementation of the scheme, the system comprises a content acquisition module, a data preprocessing module, a monitoring analysis module and an encryption transmission module; the invention realizes the efficient and safe transmission of video data and improves the working efficiency of video monitoring.
The foregoing has shown and described the basic principles, principal features and advantages of the invention. It will be understood by those skilled in the art that the present invention is not limited to the above-described embodiments, and that the above-described embodiments and descriptions are only preferred embodiments of the present invention, and are not intended to limit the invention, and that various changes and modifications may be made therein without departing from the spirit and scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (2)
1. The video monitoring safety management method based on the Internet of things is characterized in that video monitoring equipment is connected with an Internet of things platform through the Internet of things, and the video monitoring safety management method based on the Internet of things comprises the following steps:
monitoring content acquisition is carried out through video monitoring equipment to obtain first video monitoring data, and the first video monitoring data are stored into an IP-SAN;
reading the first video monitoring data in the IP-SAN, transmitting the first video monitoring data to an Internet of things platform, and carrying out data preprocessing on the first video monitoring data to obtain second video monitoring data;
Determining an application scene of current video monitoring, constructing a corresponding monitoring processing model, inputting the second video monitoring data into the monitoring processing model, and outputting a monitoring analysis result through the monitoring processing model;
Integrating the monitoring analysis results to generate target video monitoring content and corresponding safety early warning information, and encrypting and transmitting the target video monitoring content to a user terminal based on the safety early warning information;
performing data preprocessing on the first video monitoring data to obtain second video monitoring data, wherein the data preprocessing comprises the following steps:
Acquiring image data in the first video monitoring data, and determining that two 3×3 matrix operators are respectively used for the image data in the transverse and longitudinal directions of the first video monitoring data;
Performing first-order convolution operation on the matrix operator in the transverse and longitudinal directions and pixel points on the image to obtain a gradient derivative in the transverse and longitudinal directions;
adding the first step derivative in the transverse and longitudinal directions and solving the root mean square to obtain gradient data in the transverse and longitudinal directions;
combining the gradient data in the transverse and longitudinal directions, comparing the gradient data with a set threshold value, determining whether the data of each pixel position is completely black or completely white, and outputting first image data;
And combining the gradient data in the transverse and longitudinal directions, comparing the gradient data with a set threshold value, determining whether the data of each pixel position is completely black or completely white, and outputting first image data, wherein the method further comprises the following steps:
Acquiring the first image data, and extracting sub-bands corresponding to each frequency band of the first image data to obtain a high-frequency band sub-band and a low-frequency band sub-band;
Blocking the high-frequency sub-band and the low-frequency sub-band, and calculating image noise data based on a partial differential equation;
carrying out dimension reduction processing on the first image data according to the image noise data, and connecting noise singular points of all the block areas in the image noise data to form a linear structure;
obtaining reconstruction data of the first image data after dimension reduction to obtain the first image data, and carrying out image enhancement on the first image data by adopting wavelet transformation and a mathematical morphology method to obtain second image data;
integrating the second image data to obtain second video monitoring data;
the determining the application scene of the current video monitoring and constructing a corresponding monitoring processing model comprise the following steps:
Acquiring a history monitoring image of a video monitoring application scene, forming a training set, and inputting the training set into a YOLOv network structure;
adopting a step-by-step channel convolution mode, in a two-dimensional plane, the convolution kernels are in one-to-one correspondence with the number of channels, one convolution kernel processes one channel, and 3 feature images can be generated after the history monitoring images in the training set are processed;
the number of the feature mappings subjected to the depth convolution is consistent with the channel value of the input layer, and the feature combination is performed by adopting point-by-point convolution to obtain new feature mappings;
performing model adjustment by adopting sparse training and pruning technology to obtain a monitoring processing model;
The model adjustment is carried out by adopting sparse training and pruning technology to obtain a monitoring processing model, which comprises the following steps:
Pruning is carried out on each regularization module as a whole, pruning is carried out according to the sequence from low to high of the weight average value of the regularization module, and pruning is carried out according to the ratio of one half of the regularization module in the primary pruning process;
compensating by combining with the fine adjustment of the model, and continuously carrying out cyclic iteration to obtain a network structure after layer pruning;
Determining a channel pruning threshold corresponding to the global pruning rate according to the pruning parameters and the scaling sparseness of the regularization module, and determining a protection threshold of each neural network layer according to the channel protection threshold parameters;
When the pruning threshold value is larger than the pruning threshold value and smaller than the channel protection threshold value are simultaneously established, channel pruning is carried out on the network structure after the layer pruning;
Integrating the monitoring analysis results to generate target video monitoring content and corresponding safety early warning information, wherein the method comprises the following steps:
Acquiring a video stream of the second video monitoring data, acquiring a video frame image from the video stream, and correcting the video frame image by adopting a division model;
Normalizing the corrected video frame image, taking the normalized video frame image as the input of Deeplabv & lt3+ & gt model, extracting the characteristics layer by layer through a coding network, and obtaining the characteristic codes of the pictures;
Upsampling the feature codes through a decoding network to obtain a segmentation mask diagram with the same resolution as the image input by the model;
video stitching fusion is carried out based on the segmentation mask diagram to obtain target video monitoring content, and safety early warning information is obtained according to the monitoring analysis result;
the encrypting and transmitting the target video monitoring content to the user terminal based on the safety precaution information comprises the following steps:
receiving the target video monitoring, and extracting a header field and a content field of an original video data packet of the target video monitoring;
performing hash operation on the header field and the content field respectively to obtain a corresponding hash value field;
Forming a new video data packet by the hash value field corresponding to the original video data packet and the content field of the original video data packet;
and obtaining a new data stream of the target video monitoring according to the new video data packet, and encrypting the new data stream of the target video monitoring by adopting a cross chaos algorithm.
2. The video monitoring safety management system based on the Internet of things is characterized by comprising a content acquisition module, a data preprocessing module, a monitoring analysis module and an encryption transmission module, wherein the content acquisition module is used for acquiring monitoring content through video monitoring equipment to obtain first video monitoring data, and storing the first video monitoring data into an IP-SAN;
The data preprocessing module is used for reading the first video monitoring data in the IP-SAN, transmitting the first video monitoring data to an Internet of things platform, and preprocessing the first video monitoring data to obtain second video monitoring data;
The monitoring analysis module is used for determining an application scene of current video monitoring, constructing a corresponding monitoring processing model, inputting the second video monitoring data into the monitoring processing model, and outputting a monitoring analysis result through the monitoring processing model, wherein a historical monitoring image of the video monitoring application scene is obtained, a training set is formed, and the training set is input into a YOLOv network structure; adopting a step-by-step channel convolution mode, in a two-dimensional plane, the convolution kernels are in one-to-one correspondence with the number of channels, one convolution kernel processes one channel, and 3 feature images can be generated after the history monitoring images in the training set are processed; the number of the feature mappings subjected to the depth convolution is consistent with the channel value of the input layer, and the feature combination is performed by adopting point-by-point convolution to obtain new feature mappings; performing model adjustment by adopting sparse training and pruning technology to obtain a monitoring processing model;
The model adjustment is carried out by adopting sparse training and pruning technology to obtain a monitoring processing model, which comprises the following steps: pruning is carried out on each regularization module as a whole, pruning is carried out according to the sequence from low to high of the weight average value of the regularization module, and pruning is carried out according to the ratio of one half of the regularization module in the primary pruning process; compensating by combining with the fine adjustment of the model, and continuously carrying out cyclic iteration to obtain a network structure after layer pruning; determining a channel pruning threshold corresponding to the global pruning rate according to the pruning parameters and the scaling sparseness of the regularization module, and determining a protection threshold of each neural network layer according to the channel protection threshold parameters; when the pruning threshold value is larger than the pruning threshold value and smaller than the channel protection threshold value are simultaneously established, channel pruning is carried out on the network structure after the layer pruning;
The encryption transmission module is used for integrating the monitoring analysis results to generate target video monitoring content and corresponding safety early warning information, encrypting and transmitting the target video monitoring content to a user terminal based on the safety early warning information, wherein a video stream of the second video monitoring data is acquired, video frame images are acquired from the video stream, and a division model is adopted to correct the video frame images; normalizing the corrected video frame image, taking the normalized video frame image as the input of Deeplabv & lt3+ & gt model, extracting the characteristics layer by layer through a coding network, and obtaining the characteristic codes of the pictures; upsampling the feature codes through a decoding network to obtain a segmentation mask diagram with the same resolution as the image input by the model; video stitching fusion is carried out based on the segmentation mask diagram to obtain target video monitoring content, and safety early warning information is obtained according to the monitoring analysis result;
The content acquisition module comprises an acquisition sub-module, a convolution operation sub-module, a calculation sub-module and a data combination sub-module, wherein the acquisition sub-module is used for acquiring image data in the first video monitoring data and determining that two 3X 3 matrix operators are respectively used for the image data in the transverse and longitudinal directions of the first video monitoring data;
The convolution operation sub-module is used for carrying out first-order convolution operation on the matrix operator in the transverse and longitudinal directions and the pixel points on the image to obtain a step derivative in the transverse and longitudinal directions;
the computing sub-module is used for adding the first step derivative in the transverse and longitudinal directions and solving the root mean square to obtain gradient data in the transverse and longitudinal directions;
The data combination sub-module is used for combining the gradient data in the transverse and longitudinal directions, comparing the gradient data with a set threshold value, determining whether the data of each pixel position is completely black or completely white, and outputting first image data;
The encryption transmission module comprises an extraction sub-module, a hash operation sub-module, a composition sub-module and an encryption sub-module, wherein the extraction sub-module is used for receiving the target video monitoring and extracting the head field and the content field of an original video data packet of the target video monitoring;
The hash operation sub-module is used for carrying out hash operation on the header field and the content field respectively to obtain a corresponding hash value field;
the composition sub-module is used for composing the hash value field corresponding to the original video data packet and the content field of the original video data packet into a new video data packet;
And the encryption sub-module is used for obtaining the new data stream of the target video monitoring according to the new video data packet, and encrypting the new data stream of the target video monitoring by adopting a cross chaos algorithm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311308817.0A CN117061711B (en) | 2023-10-11 | Video monitoring safety management method and system based on Internet of things |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311308817.0A CN117061711B (en) | 2023-10-11 | Video monitoring safety management method and system based on Internet of things |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117061711A CN117061711A (en) | 2023-11-14 |
| CN117061711B true CN117061711B (en) | 2024-07-09 |
Family
ID=
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115086718A (en) * | 2022-07-19 | 2022-09-20 | 广州万协通信息技术有限公司 | Video stream encryption method and device |
| CN116193075A (en) * | 2023-01-18 | 2023-05-30 | 国网浙江省电力有限公司海盐县供电公司 | Intelligent monitoring method and system based on control of Internet of things |
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115086718A (en) * | 2022-07-19 | 2022-09-20 | 广州万协通信息技术有限公司 | Video stream encryption method and device |
| CN116193075A (en) * | 2023-01-18 | 2023-05-30 | 国网浙江省电力有限公司海盐县供电公司 | Intelligent monitoring method and system based on control of Internet of things |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Thai et al. | Statistical model of quantized DCT coefficients: Application in the steganalysis of Jsteg algorithm | |
| Soliman et al. | Efficient HEVC steganography approach based on audio compression and encryption in QFFT domain for secure multimedia communication | |
| Narayanaperumal et al. | VLSI Implementations of Compressive Image Acquisition using Block Based Compression Algorithm. | |
| Stamm et al. | Wavelet-based image compression anti-forensics | |
| CN111050022B (en) | High-security image transmission system and method | |
| Zhang et al. | Harnessing the hybrid cloud for secure big image data service | |
| CN115063492B (en) | Method for generating countermeasure sample for resisting JPEG compression | |
| US20210225039A1 (en) | Predictive Joint Compression and Encryption for Images and Videos | |
| CN111369548A (en) | No-reference video quality evaluation method and device based on generation countermeasure network | |
| Prabakaran et al. | Dual transform based steganography using wavelet families and statistical methods | |
| Li et al. | Medical images lossless recovery based on POB number system and image compression | |
| Zafari et al. | Attention-based generative neural image compression on solar dynamics observatory | |
| Chai et al. | CSENMT: A deep image compressed sensing encryption network via multi-color space and texture feature | |
| CN117061711B (en) | Video monitoring safety management method and system based on Internet of things | |
| Vidhya et al. | Evaluation and performance analysis of Chinese remainder theorem and its application to lossless image compression | |
| Wang et al. | A customized deep network based encryption-then-lossy-compression scheme of color images achieving arbitrary compression ratios | |
| Kekre et al. | Image Reconstruction using Fast Inverse Half tone and Huffman Coding Technique | |
| CN117061711A (en) | Video monitoring safety management method and system based on Internet of things | |
| Cemiloglu et al. | Blind video quality assessment via spatiotemporal statistical analysis of adaptive cube size 3D‐DCT coefficients | |
| Bairi et al. | PSCS-Net: Perception optimized image reconstruction network for autonomous driving systems | |
| Kumar et al. | A novel method for image compression using spectrum | |
| Ayyoubzadeh et al. | Lossless compression of mosaic images with convolutional neural network prediction | |
| Mydin et al. | A Study on Medical Image Compression Techniques based on Huffman Coding and Discrete Wavelet Transform (DWT) | |
| El-Shahed et al. | Video Steganography using 3D Stationary Wavelet Transform | |
| Rajakumar et al. | Lossy image compression using multiwavelet transform for wireless transmission |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |