CN113065596B - Industrial safety real-time monitoring system based on video analysis and artificial intelligence - Google Patents
Industrial safety real-time monitoring system based on video analysis and artificial intelligence Download PDFInfo
- Publication number
- CN113065596B CN113065596B CN202110360777.9A CN202110360777A CN113065596B CN 113065596 B CN113065596 B CN 113065596B CN 202110360777 A CN202110360777 A CN 202110360777A CN 113065596 B CN113065596 B CN 113065596B
- Authority
- CN
- China
- Prior art keywords
- convex hull
- shape
- weight
- image
- safety
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to the field of artificial intelligence, in particular to an industrial safety real-time monitoring system based on video analysis and artificial intelligence. The system comprises an image acquisition module, a processing module and a display module, wherein the image acquisition module is used for acquiring the shape of a convex hull to be detected corresponding to a mechanical image; the working mode acquisition module is used for obtaining a variance matrix according to the deviation of the shape of each convex hull and the average shape, calculating eigenvectors and eigenvalues of the variance matrix and acquiring n working modes corresponding to the first n eigenvalues with the largest median of the eigenvalues; the weight vector acquisition module is used for acquiring weight vectors according to n eigenvectors corresponding to the first n eigenvalues, the average shape and the shape of the convex hull to be detected; and the safety evaluation index acquisition module is used for acquiring the safety evaluation index according to each weight in the weight vector. The system solves the technical problem that misjudgment is caused by measuring the safety risk of the machine only by comparing the similarity between the current mechanical behavior and the standard behavior of the current mechanical behavior, and improves the accuracy and reliability of the mechanical safety evaluation index.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an industrial safety real-time monitoring system based on video analysis and artificial intelligence.
Background
Industrial automation is a process that widely uses automatic control and automatic adjustment devices in industrial production to replace manual operation machines and mechanical systems for processing. In the process, the assistance of a machine cannot be separated no matter the machine is semi-automatic or fully automatic, however, the continuous and repeated automatic production of the machine can cause inevitable abrasion and mechanical failure, and potential safety hazards exist.
In the prior art, the safety of the machine is detected by comparing the similarity between the current behavior and the standard behavior of the machine to measure the safety index of the machine, however, the machine is a continuous process in the actual operation process, so that the technical problem of misjudgment and misjudgment can exist when the safety risk of the machine is judged only by the similarity between the current behavior and the standard behavior.
Disclosure of Invention
In order to solve the above technical problems, the present invention aims to provide an industrial safety real-time monitoring system based on video analysis and artificial intelligence, and the adopted technical scheme is as follows:
the embodiment of the invention provides an industrial safety real-time monitoring system based on video analysis and artificial intelligence, which comprises:
the image acquisition module is used for acquiring a key point image from the mechanical image and performing convex hull detection on the key point image to obtain the shape of a convex hull to be detected;
the working mode acquisition module is used for acquiring an average shape according to a plurality of convex hull shapes in historical data, acquiring a variance matrix according to the deviation of each convex hull shape and the average shape, calculating eigenvectors and corresponding eigenvalues of the variance matrix, and acquiring n working modes corresponding to the first n eigenvalues with the largest median of the eigenvalues;
a weight vector obtaining module, configured to obtain weight vectors of n working modes corresponding to the shape of the convex hull to be detected according to the n eigenvectors corresponding to the first n eigenvalues, the average shape, and the shape of the convex hull to be detected;
a safety assessment index obtaining module, configured to determine whether each weight in the weight vector is within a corresponding threshold range, and if at least one weight is not within the corresponding threshold range, the safety assessment index is zero; and if the weight of each weight is within the corresponding threshold value range, acquiring a safety evaluation index according to each weight.
Further, the industrial safety real-time monitoring device further comprises an image storage module, which is used for setting the number of storage layers of the bit planes corresponding to the mechanical images according to the safety evaluation index.
Further, the calculation formula of the average shape is as follows:
wherein the content of the first and second substances,
for the average shape, N is the number of convex hull shapes in the history data, xiIn the shape of the ith convex hull.
Further, the formula of the weight vector is as follows:
wherein b ═ b1,b2,…,bm,…,bn) B is the weight vector, bmFor the weight corresponding to the mth operation mode, P ═ P1,p2,…,pm,…,pn)，pmAnd the characteristic vector is corresponding to the mth working mode, and x is the shape of the convex hull to be detected.
Further, the security assessment index acquisition module includes:
a convex hull shape set obtaining unit, configured to obtain a convex hull shape set corresponding to each operating mode from the plurality of convex hull shapes;
a threshold value obtaining unit for obtaining a threshold value range corresponding to each weight
σmThe variance of the convex hull shape set corresponding to the mth working mode;
a weight sorting unit for sorting the weight vectors from large to small to obtain a weight sequence (d)1,d2,…,dn)。
Further, the formula of the safety assessment index is as follows:
wherein S is the safety assessment index.
Further, the image storage module includes:
the discrete cosine transform unit is used for performing discrete cosine transform on the gray level image corresponding to the mechanical image to obtain frequency data;
the bit layering processing unit is used for converting the gray value in the gray image into a binary system, each binary value corresponds to one bit plane, and the number of storage layers of the gray image is set according to the safety evaluation index;
and the frequency data processing unit is used for updating the frequency data according to the data obtained by the number of the storage layers.
Further, when the safety evaluation index is smaller than the minimum threshold value, the mechanical image is directly stored without discrete cosine transform and bit layering processing.
Further, when the safety evaluation index is larger than a minimum threshold, the number of storage layers is obtained according to the gray value range of the gray image and the safety evaluation index.
The embodiment of the invention at least has the following beneficial effects:
according to the embodiment of the invention, the convex hull shape corresponding to the mechanical image is obtained through the image obtaining module; obtaining m most main working modes of the machine in the working process through a working mode obtaining module; obtaining weight vectors corresponding to the convex hull shape and n working modes of the machine through a weight vector obtaining module; the safety assessment index acquisition module acquires the safety assessment index according to each weight in the weight vector, so that the technical problem that misjudgment is caused by the fact that the safety risk of the machine is measured only by comparing the similarity between the current mechanical behavior and the standard behavior in the prior art is solved, and the accuracy and the reliability of the safety assessment index of the machine are improved.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions and advantages of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts;
fig. 1 is a block diagram of an industrial safety real-time monitoring system based on video analysis and artificial intelligence according to an embodiment of the present invention;
fig. 2 is a schematic diagram illustrating a key point annotation in a key point detection network according to an embodiment of the present invention;
fig. 3 is a block diagram of an industrial safety real-time monitoring system based on video analysis and artificial intelligence according to another embodiment of the present invention.
Detailed Description
In order to further illustrate the technical means and effects of the present invention adopted to achieve the predetermined objects, the following detailed description, the structure, the features and the effects of the industrial safety real-time monitoring system based on video analysis and artificial intelligence according to the present invention are provided with the accompanying drawings and the preferred embodiments. In the following description, different "one embodiment" or "another embodiment" refers to not necessarily the same embodiment. Furthermore, the particular features, structures, or characteristics may be combined in any suitable manner in one or more embodiments.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following describes a specific scheme of the industrial safety real-time monitoring system based on video analysis and artificial intelligence in detail with reference to the accompanying drawings.
Referring to fig. 1, a block diagram of an industrial safety real-time monitoring system based on video analysis and artificial intelligence according to an embodiment of the present invention is shown, where the industrial safety real-time monitoring system 100 includes:
and the image acquisition module 10 is configured to acquire a key point image from the mechanical image, and perform convex hull detection on the key point image to obtain a shape of a convex hull to be detected.
The working mode obtaining module 20 is configured to obtain an average shape according to a plurality of convex hull shapes in the historical data, obtain a variance matrix according to a deviation between each convex hull shape and the average shape, calculate eigenvectors and corresponding eigenvalues of the variance matrix, and obtain n working modes corresponding to the first n eigenvalues with the largest median of the eigenvalues.
The weight vector obtaining module 30 is configured to obtain weight vectors of n working modes corresponding to the shape of the convex hull to be detected according to the n eigenvectors corresponding to the first n eigenvalues, the average shape, and the shape of the convex hull to be detected.
A safety assessment index obtaining module 40, configured to determine whether each weight in the weight vector is within a corresponding threshold range, and if at least one weight is not within the corresponding threshold range, the safety assessment index is zero; and if each weight is within the corresponding threshold value range, acquiring a safety evaluation index according to each weight.
In summary, in the embodiment of the present invention, the image obtaining module 10 obtains the convex hull shape corresponding to the mechanical image; obtaining the most main n working modes of the machine in the working process through the working mode obtaining module 20; obtaining weight vectors corresponding to the convex hull shape and n working modes of the machine through a weight vector obtaining module 30; the safety assessment index acquisition module 40 acquires the safety assessment index according to each weight in the weight vector, so that the technical problem that misjudgment is caused by the fact that the safety risk of the machine is measured only by comparing the similarity between the current mechanical behavior and the standard behavior in the prior art is solved, and the accuracy and the reliability of the safety assessment index of the machine are improved.
Preferably, the image acquisition module 10 in this embodiment includes a key point detection network 101.
And the key point detection network 101 is used for inputting the mechanical image acquired by the monitoring equipment into the key point detection network to obtain a key point image.
The key point detection network in this embodiment adopts a DNN network with an Encoder-Decoder (Encoder-Decoder) structure, and the specific training process is as follows:
(1) referring to fig. 2, joint points of the mechanical arm in the sample image data set are labeled as key points 1 to 10, and gaussian kernel convolution is performed on ten key points to obtain tag data.
The sample image dataset is a collection of multiple mechanical images acquired by the monitoring device. 80% of the sample image data sets were training sets, and the remaining 20% were validation sets.
(2) End-to-end training is performed on the mechanical image and corresponding label data.
(3) And optimizing network parameters by adopting a mean square error loss function.
Preferably, the operation mode acquisition module 20 in the present embodiment includes an average shape acquisition unit 201, a variance matrix acquisition unit 202, and an operation mode acquisition unit 203.
An average shape obtaining unit 201, configured to perform similarity transformation on a plurality of convex hull shapes in the history data and align the convex hull shapes to obtain an average shape.
The alignment process comprises the following steps:
(1) and randomly selecting one convex hull shape from the historical data as a reference shape, and aligning the pose shape in each of the rest convex hull shapes with the reference shape in sequence through similarity transformation.
(2) Calculating the average shape of all convex hull shapes after alignment, and aligning the average shape to the first convex hull shape through rotating, scaling and translating operations, namely, specifying the size, the direction and the origin of the average shape.
(3) Each convex hull shape in the training set of convex hull shapes is aligned with the average shape.
(4) And (3) repeating the steps (2) and (3) until convergence, and judging whether the convergence is achieved by calculating the average distance between each convex hull shape and the average shape.
The formula for the average shape is as follows:
wherein the content of the first and second substances,
is the average shape, N is the number of convex hull shapes in the history, xiIn the shape of the ith convex hull.
A variance matrix obtaining unit 202 for calculating a deviation between each convex hull shape and the average shape, and obtaining a variance matrix for each convex hull shape according to the deviation values.
Deviation dxiThe formula of (1) is:
the formula of the variance matrix a is:
and the working mode acquiring unit 203 is used for acquiring the working mode of the machine according to the eigenvalue of the variance matrix. The method comprises the following specific steps:
(1) let the eigenvector of the variance matrix be piThen the corresponding eigenvalue can be obtained according to the definition of the matrix eigenvector.
Definition of feature vectors:
Api＝γipi
wherein, γiAs feature vectors piThe corresponding characteristic value.
(2) Normalizing the feature vectors, i.e.
pi Tpi＝1
(3) And acquiring the maximum n eigenvalues from all eigenvalues of the variance matrix, wherein the working mode corresponding to the n eigenvalues is the main mechanical working mode.
Since most changes of the machine can be obtained through the comprehensive action of few changes, the number of the main working modes can be determined to represent most action behaviors of the machine.
For machines with different division of work in different environments, the corresponding main operating modes are different, and in the embodiment, the machines have four different operating modes in common, that is, n is 4. In other embodiments, the number of mechanical modes of operation may be selected by the practitioner based on the circumstances.
Preferably, the weight vector obtaining module 30 in this embodiment is configured to obtain a weight vector of each working mode corresponding to the shape of the convex hull. The method comprises the following specific steps:
since any convex hull shape can be obtained by weighted sum of the average shape and the deviation of the convex hull shape and the four working modes, the convex hull shape x to be measured is
Wherein b ═ b1,b2,…,bm,…,bn) B is a weight vector, bmFor the weight corresponding to the mth operation mode, P ═ P1,p2,…,pm,…,pn)，pmAnd the feature vector corresponding to the mth working mode.
Since each eigenvector of the variance matrix is orthogonal, the weight vector is calculated as:
and substituting the n eigenvectors corresponding to the first n eigenvalues, the average shape and the convex hull shape to be measured into a calculation formula of the weight vector to obtain the weight vector corresponding to the convex hull shape to be measured and related to the n working modes.
Preferably, the security assessment index acquisition module 40 in this embodiment includes a convex hull shape set acquisition unit 401, a threshold acquisition unit 402, a weight sorting unit 403, and a security assessment index calculation unit 404.
A convex hull shape set obtaining unit 401, configured to obtain a convex hull shape set corresponding to each operating mode from a plurality of convex hull shapes of the history data.
A threshold value obtaining unit 402, configured to obtain a threshold value range corresponding to each weight
σmThe variance of the convex hull shape set corresponding to the mth operating mode.
A weight sorting unit 403 for sorting the weight vectors from large to small to obtain a new weight sequence (d)1,d2,…,dn)。
And a safety evaluation index calculation unit 404, configured to obtain a safety evaluation index according to the weight of each working mode corresponding to the shape of the convex hull to be detected. The method comprises the following specific steps:
(1) according to the standard normal distribution curve, most convex hull shape samples are distributed in
And within the range, when at least one weight in the weight vector is not in the corresponding threshold range, the safety evaluation index is zero.
(2) And if each weight is within the corresponding threshold value range, obtaining the safety evaluation index according to a calculation formula of the safety evaluation index.
The shape of the convex hull to be measured is most similar to the feature vector with the largest weight, so that the weight of the convex hull in the safety assessment index is the largest. And because any convex hull shape is obtained by n main characteristic vectors through comprehensive changes, the similarity between the convex hull shape to be detected and the rest n-1 working modes also needs to be reflected in the safety evaluation index, and the more similar the convex hull shape to be detected and the rest n-1 working modes, the larger the integral weight ratio is.
The calculation formula of the safety evaluation index S is as follows:
the larger the value of the safety evaluation index S is, the closer the shape of the convex hull to be detected is to the normal working mode, the higher the corresponding safety index is, and the lower the risk of hidden danger is.
Referring to fig. 3, preferably, a large number of monitoring videos are obtained by a machine in a daily working process, and if the videos are directly stored, a huge memory consumption and expenditure expense are generated, so as to facilitate an enterprise to manage and maintain the videos during the machine working process, and reduce the space and cost of video storage while not affecting the detection of the machine safety, in which the industrial safety real-time monitoring system 100 in this embodiment further includes an image storage module 50 for setting the number of storage layers of bit planes corresponding to the machine images according to the safety evaluation index.
The image storage module 50 includes a discrete cosine transform unit 501, a bit layering processing unit 502, and a frequency data processing unit 503.
A discrete cosine transform unit 501, configured to perform discrete cosine transform on the grayscale image corresponding to the mechanical image to obtain frequency data.
The human eye is sensitive to information of low-frequency characteristics in the mechanical image, such as the overall brightness of the machine, and is insensitive to high-frequency detail information in the mechanical image, however, a large amount of high-frequency information irrelevant to the working mode of the machine exists in the mechanical image, so that the high-frequency information can be transmitted little or not, and only the low-frequency part is transmitted in the process of storing the mechanical image. Therefore, in the embodiment of the present invention, the grayscale image is processed by using discrete cosine transform, the mechanical image signal described in the spatial domain is transformed into the frequency domain, and then the transformed frequency coefficient is processed.
The discrete cosine transform comprises the following specific steps:
(1) the grayscale image is first decomposed into 8 × 8 sub-block images that do not overlap each other.
(2) Each sub-block image is fed into a discrete cosine transform encoder to be transformed from a spatial domain to a frequency domain.
(3) And obtaining the frequency coefficient of each transformed sub-block, wherein the low-frequency coefficient at the upper left corner concentrates a large amount of energy, and the high-frequency coefficient at the lower right corner has small energy.
The bit layering processing unit 502 is configured to convert the gray values in the gray image into binary systems, each binary value corresponds to a bit plane, and the number of storage layers of the gray image is set according to the security evaluation index.
In the frequency data obtained after discrete cosine transform, usually the larger frequency coefficient is concentrated in the upper left corner, and it is also the most important coefficient, and then gradually decreases in the direction going to the lower right corner. Therefore, the frequency data after the gray level image transformation can be coded according to the sequence of the bit planes under different transmission characteristics, namely, the bit plane with high bits is coded first, and for some small coefficients, the high data bits are zero necessarily, so that the coded data volume is reduced.
The specific steps of bit plane layering are as follows:
(1) the storage space can be saved by converting the gray scale image data into a double type.
(2) The height r and width c of the grayscale image are obtained.
(3) An 8-level r × c zero matrix is defined for storing the final results.
(4) The gray image data is converted into a binary string of 8 bits, and the binary string matrix is converted into a matrix of [ r, c,8 ].
(5) Traversing each pixel point in the gray level image, taking out the binary number of each pixel point in the matrix, traversing each bit value of the binary number, judging whether the binary number is 1, and if the binary number is 1, assigning the value of the corresponding bit plane to be 1; otherwise, assigning the value of the corresponding bit plane as 0; and finally outputting the image of each layer.
Since each bit plane records certain information of the grayscale image, and the importance of the information contained in the bit planes from the high level to the low level is lower and lower, the number of storage layers of the bit planes is determined according to the mechanical security evaluation index degree in this embodiment, and the specific steps are as follows:
(1) the minimum threshold value of the safety evaluation index is set to be S ═ 0.3, namely when the safety evaluation index S is lower than 0.3, the system timely sends out early warning information, and mechanical images at the stage are shot and stored in real time in the whole process, and discrete cosine transform and bit plane layering processing are not carried out any more.
(2) And when the safety evaluation index is larger than the minimum threshold, determining the storage layer number of the bit plane according to the size of the safety evaluation index. The method specifically comprises the following steps: obtaining a range of gray values [0,2 ] for a gray imagek]K has a value range of [1, 8]]If the number of the plane with the highest layer of the bit plane being pure black is 8-, the number of the storage layers of the gray image is c layers below the full black layer. The highest layer is the bit plane storing the most information, and the highest layer is the 8 th layer in this embodiment.
The number c of storage layers is:
c＝k+3-10m，c≥1
the safety evaluation index with 10m being 10 times is rounded down in the adjacent integer range, for example, when the safety evaluation index S is in the range of 0.3-0.4, 10m is 3.
A specific example of the number of storage layers is given below:
when the gray scale value range of the gray scale image is [0,127] and the security evaluation index s is 0.56, the method for calculating the number of storage layers is as follows: if the gray value range is [0,127], k is 7, the number of bit planes with the highest layer being pure black is 1, that is, the bit plane of the 8 th layer is pure black; if s is 0.56, 10m is 5, and the number c of storage layers corresponding to the grayscale image is 5, that is, only the data on the 7 th, 6 th, 5 th, 4 th, and 3 rd bit planes are stored.
And a frequency data processing unit 503, configured to update frequency data according to the data obtained by the number of storage layers, and compress and store the mechanical image according to the updated frequency data.
It should be noted that: the precedence order of the above embodiments of the present invention is only for description, and does not represent the merits of the embodiments. And specific embodiments thereof have been described above. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.
The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.
The present invention is not limited to the above preferred embodiments, and any modifications, equivalent replacements, improvements, etc. within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (8)
1. An industrial safety real-time monitoring system based on video analysis and artificial intelligence, characterized in that the system comprises:
the image acquisition module is used for acquiring a key point image of a joint point of the mechanical arm from the mechanical image and performing convex hull detection on the key point image to obtain the shape of a convex hull to be detected;
the working mode acquisition module is used for acquiring an average shape according to a plurality of convex hull shapes in historical data, acquiring a variance matrix according to the deviation of each convex hull shape and the average shape, calculating eigenvectors and corresponding eigenvalues of the variance matrix, and acquiring n working modes corresponding to the first n eigenvalues with the largest median of the eigenvalues;
a weight vector obtaining module, configured to obtain weight vectors of n working modes corresponding to the shape of the convex hull to be detected according to the n eigenvectors corresponding to the first n eigenvalues, the average shape, and the shape of the convex hull to be detected;
a safety assessment index obtaining module, configured to determine whether each weight in the weight vector is within a corresponding threshold range, and if at least one weight is not within the corresponding threshold range, the safety assessment index is zero; if each weight is within the corresponding threshold value range, obtaining a safety evaluation index according to each weight;
the formula of the weight vector is as follows:
wherein b ═ b1,b2,…,bm,…,bn) B is the weight vector, bmFor the weight corresponding to the mth operation mode, P ═ P1,p2,…,pm,…,pn)，pmIs the characteristic vector corresponding to the mth working mode, x is the shape of the convex hull to be detected,
is an average shape.
2. The system according to claim 1, wherein the industrial safety real-time monitoring system further comprises an image storage module configured to set the number of storage layers of the bit plane corresponding to the mechanical image according to the safety assessment indicator.
3. The system for real-time monitoring of industrial safety based on video analysis and artificial intelligence as claimed in claim 1, wherein the calculation formula of the average shape is as follows:
wherein the content of the first and second substances,
for the average shape, N is the number of convex hull shapes in the history data, xiIn the shape of the ith convex hull.
4. The industrial safety real-time monitoring system based on video analysis and artificial intelligence as claimed in claim 1, wherein the safety assessment index obtaining module comprises:
a convex hull shape set obtaining unit, configured to obtain a convex hull shape set corresponding to each operating mode from the plurality of convex hull shapes;
a threshold value obtaining unit for obtaining a threshold value range corresponding to each weight
σmThe variance of the convex hull shape set corresponding to the mth working mode;
a weight sorting unit for sorting the weight vectors from large to small to obtain a weight sequence (d)1,d2,…,dn)。
5. The industrial safety real-time monitoring system based on video analysis and artificial intelligence is characterized in that the safety assessment index is formulated as follows:
wherein S is the safety assessment index.
6. The industrial safety real-time monitoring system based on video analysis and artificial intelligence as claimed in claim 1, wherein the image storage module comprises:
the discrete cosine transform unit is used for performing discrete cosine transform on the gray level image corresponding to the mechanical image to obtain frequency data;
the bit layering processing unit is used for converting the gray value in the gray image into a binary system, each binary value corresponds to one bit plane, and the number of storage layers of the gray image is set according to the safety evaluation index;
and the frequency data processing unit is used for updating the frequency data according to the data obtained by the number of the storage layers.
7. The industrial safety real-time monitoring system based on video analysis and artificial intelligence as claimed in claim 6, wherein when the safety assessment index is smaller than a minimum threshold, the mechanical image is directly stored without discrete cosine transform and bit layering.
8. The system according to claim 6, wherein the number of storage layers is obtained according to the gray scale value range of the gray scale image and the safety assessment index when the safety assessment index is greater than a minimum threshold.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110360777.9A CN113065596B (en) | 2021-04-02 | 2021-04-02 | Industrial safety real-time monitoring system based on video analysis and artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110360777.9A CN113065596B (en) | 2021-04-02 | 2021-04-02 | Industrial safety real-time monitoring system based on video analysis and artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113065596A CN113065596A (en) | 2021-07-02 |
| CN113065596B true CN113065596B (en) | 2022-01-28 |
Family
ID=76565473
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110360777.9A Active CN113065596B (en) | 2021-04-02 | 2021-04-02 | Industrial safety real-time monitoring system based on video analysis and artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113065596B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108458668A (en) * | 2018-01-05 | 2018-08-28 | 燕山大学 | Slab edge and Head and Tail Shape automatic checkout system based on binocular vision and method |
| CN109919471A (en) * | 2019-02-27 | 2019-06-21 | 河南鑫安利安全科技股份有限公司 | A kind of wisdom air control appraisal procedure and system |
| CN110276528A (en) * | 2019-05-29 | 2019-09-24 | 上海电力学院 | A kind of control system of power plant information security risk evaluation system |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103426164B (en) * | 2013-06-09 | 2016-01-13 | 大连海事大学 | Based on scallop size computing method and the scallop sorting system of Opencv graphical analysis |
| WO2015123646A1 (en) * | 2014-02-14 | 2015-08-20 | Nant Holdings Ip, Llc | Edge-based recognition, systems and methods |
| CN107014818B (en) * | 2017-02-28 | 2020-08-04 | 深圳市维图视技术有限公司 | Laser etching defect visual detection system and method |
| CN107315410B (en) * | 2017-06-16 | 2020-05-29 | 江苏科技大学 | Automatic obstacle removing method for robot |
| CN107563432A (en) * | 2017-08-29 | 2018-01-09 | 武汉交通职业学院 | A kind of robot multiple-objective recognition methods of view-based access control model shape |
| CN107877517B (en) * | 2017-11-16 | 2021-03-30 | 哈尔滨工业大学 | Motion mapping method based on cyberporce remote operation mechanical arm |
| CN109191761A (en) * | 2018-09-06 | 2019-01-11 | 中国石油大学(华东) | A kind of method for recognizing fire disaster based on flame multiple features fusion |
| CN112541897A (en) * | 2020-12-14 | 2021-03-23 | 范会笑 | Facial paralysis degree evaluation system based on artificial intelligence |
-
2021
- 2021-04-02 CN CN202110360777.9A patent/CN113065596B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108458668A (en) * | 2018-01-05 | 2018-08-28 | 燕山大学 | Slab edge and Head and Tail Shape automatic checkout system based on binocular vision and method |
| CN109919471A (en) * | 2019-02-27 | 2019-06-21 | 河南鑫安利安全科技股份有限公司 | A kind of wisdom air control appraisal procedure and system |
| CN110276528A (en) * | 2019-05-29 | 2019-09-24 | 上海电力学院 | A kind of control system of power plant information security risk evaluation system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113065596A (en) | 2021-07-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113569903A (en) | Method, system, equipment, medium and terminal for predicting abrasion of numerical control machine tool cutter | |
| CN114494222B (en) | Vision Transformer-based intelligent fault identification method for rolling bearing | |
| CN112819080B (en) | High-precision universal three-dimensional point cloud identification method | |
| CN117150282B (en) | Secondhand equipment recycling evaluation method and system based on prediction model | |
| CN115331172A (en) | Workshop dangerous behavior recognition alarm method and system based on monitoring video | |
| CN116680643A (en) | Abnormality detection method combining deep contrast learning and density clustering | |
| CN113065596B (en) | Industrial safety real-time monitoring system based on video analysis and artificial intelligence | |
| CN113762151A (en) | Fault data processing method and system and fault prediction method | |
| CN117763494A (en) | Bearing diagnosis method and system based on multi-mode and multi-scale fusion network | |
| CN117195114A (en) | Chemical production line identification method and system | |
| CN116029617B (en) | Quality acceptance form generation method, device, equipment and readable storage medium | |
| CN117011280A (en) | 3D printed concrete wall quality monitoring method and system based on point cloud segmentation | |
| CN111639212A (en) | Image retrieval method in mining intelligent video analysis | |
| CN117079005A (en) | Optical cable fault monitoring method, system, device and readable storage medium | |
| CN116383747A (en) | Anomaly detection method for generating countermeasure network based on multi-time scale depth convolution | |
| CN112836570B (en) | Equipment abnormality detection method utilizing Gaussian noise | |
| CN114782822A (en) | Method and device for detecting state of power equipment, electronic equipment and storage medium | |
| CN114926702A (en) | Small sample image classification method based on depth attention measurement | |
| CN108710920B (en) | Indicator diagram identification method and device | |
| CN111798376A (en) | Image recognition method and device, electronic equipment and storage medium | |
| CN118010848B (en) | Intelligent anchorage device ponding detection method and system | |
| CN117636075B (en) | Special equipment instrument identification system based on computer vision | |
| CN117541583B (en) | Drill rod corrosion monitoring method and system for explosion-proof monitoring system | |
| CN118094436A (en) | Special fault-sensitive self-supervision equipment abnormality detection method and system | |
| US7133854B2 (en) | Method and circuits for encoding an input pattern using a normalizer and a classifier |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20230407 Address after: 450000 Room 1001, 1st Floor, Building 59, Headquarters Enterprise Base, No.1 Cuizhu Street, High tech Zone, Zhengzhou City, Henan Province, China, and Room 101-102, 1st Floor, Building 76, China Patentee after: Henan xin'anli Testing Technology Co.,Ltd. Address before: 100028 2905, 25 / F, building 1, yard a 6, Shuguang Sili, Chaoyang District, Beijing Patentee before: XIN'ANLIZHONG (BEIJING) TECHNOLOGY Co.,Ltd. |