CN117392594A - Video intelligent monitoring system and method for coal conveying trestle - Google Patents

Video intelligent monitoring system and method for coal conveying trestle Download PDF

Info

Publication number
CN117392594A
CN117392594A CN202310967424.4A CN202310967424A CN117392594A CN 117392594 A CN117392594 A CN 117392594A CN 202310967424 A CN202310967424 A CN 202310967424A CN 117392594 A CN117392594 A CN 117392594A
Authority
CN
China
Prior art keywords
image
preset
identification result
fault
neural network
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202310967424.4A
Other languages
Chinese (zh)
Inventor
孙泽平
徐海
单军
赵菲
沈强
尹浩洁
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huaneng Linyi Power Generation Co Ltd
Original Assignee
Huaneng Linyi Power Generation Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huaneng Linyi Power Generation Co Ltd filed Critical Huaneng Linyi Power Generation Co Ltd
Priority to CN202310967424.4A priority Critical patent/CN117392594A/en
Publication of CN117392594A publication Critical patent/CN117392594A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/30—Noise filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10016—Video; Image sequence
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30232—Surveillance

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Software Systems (AREA)
  • Multimedia (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Quality & Reliability (AREA)
  • Image Analysis (AREA)

Abstract

The invention relates to the technical field of video monitoring, in particular to a video intelligent monitoring system and method of a coal conveying trestle, wherein the system comprises the following steps: the acquisition module is used for acquiring video images of the monitoring area in real time, and preprocessing the video images to obtain images to be detected; the identification module is used for inputting the image to be detected into a trained neural network model, and the neural network model performs feature extraction and identification on the image to be detected to obtain an identification result; the judging module is used for calculating the confidence coefficient of the identification result, judging whether the identification result is in a fault feature library or not if the confidence coefficient is not smaller than a preset confidence coefficient threshold value, and sending an alarm signal if the identification result is in the fault feature library; the invention solves the problems that the neural network model has low accuracy in identifying the coal conveying image sample, and the identification result is not confirmed, so that an error alarm signal is sent, and the working efficiency of the coal conveying system is reduced.

Description

Video intelligent monitoring system and method for coal conveying trestle
Technical Field
The invention relates to the technical field of video monitoring, in particular to a video intelligent monitoring system and method of a coal conveying trestle.
Background
The application of the artificial intelligent monitoring system in the thermal power plant lays a foundation for realizing the intelligent power plant. Along with popularization and application of advanced video identification technology, intelligent monitoring of a power plant coal conveying system can gradually reach a standardized and normalized mode.
In the prior art, the industrial television system is optimized and reformed by an artificial intelligence technology, the front-end high-definition camera is combined with the neural network model to replace manual inspection work, but the neural network model has low accuracy in identifying the coal conveying image sample, and the identification result is not confirmed, so that an error alarm signal is sent, and the working efficiency of the coal conveying system is reduced.
Disclosure of Invention
Aiming at the problems in the prior art, the invention aims to provide a video intelligent monitoring system and method for a coal conveying trestle, which solve the problems that the neural network model has low accuracy in identifying coal conveying image samples, and the identification result is not confirmed, so that an error alarm signal is sent.
In order to achieve the above purpose, the invention provides a video intelligent monitoring system and method for a coal conveying trestle, wherein the system comprises:
the acquisition module is used for acquiring video images of the monitoring area in real time, and preprocessing the video images to obtain images to be detected;
the identification module is used for inputting the image to be detected into a trained neural network model, and the neural network model performs feature extraction and identification on the image to be detected to obtain an identification result;
the judging module is used for calculating the confidence coefficient of the identification result, judging whether the identification result is in a fault feature library or not if the confidence coefficient is not smaller than a preset confidence coefficient threshold value, and sending an alarm signal if the identification result is in the fault feature library.
In some embodiments of the present application, the trained neural network model comprises:
acquiring a historical video image, preprocessing the historical video image to obtain a historical detection image, and dividing the historical detection image into a training set and a test set, wherein the training set comprises a coal conveying stack bridge fault image and the test set comprises a coal conveying stack bridge fault image and a coal conveying stack bridge normal image;
constructing a preset neural network model, training the preset neural network model according to a training set, and extracting and identifying features of the training set by the preset neural model to obtain a first neural network model;
inputting the test set into a first neural network model to obtain a test identification result, and obtaining the application evaluation degree of the first neural model according to the test identification result;
when the application evaluation degree is not smaller than a preset application evaluation degree threshold, setting the first neural model as a trained neural network model, and when the application evaluation degree is smaller than the preset application evaluation degree threshold, adding the training set, and carrying out iterative training on the first neural network model.
In some embodiments of the present application, the feature extraction and identification of the training set by the preset neural model includes:
the preset neural model segments the historical video image in the training set, a target area is determined, the target area is extracted, an extraction result is obtained, the extraction result is matched with a fault feature library, if the extraction result is matched, the historical video image corresponding to the target area is a fault image, if the identification result is not matched, the historical video image corresponding to the target area is a normal image, and the target area is a suspected fault area;
the extraction result of the target area is as follows:
wherein,e is a preset neural model bias value, y is an activation function, and +.>For presetting a nerve model convolution kernel, h0 is a weight value of the convolution kernel, G i And (3) presetting a history video image input by a jth convolution layer in the nerve model.
In some embodiments of the present application, matching the extraction result with a fault signature library includes:
performing feature analysis on the extraction result to obtain a first preset feature set, a second preset feature set and a third preset feature set;
sequentially calculating a first preset feature set, a second preset feature set, a third preset feature set and a first matching degree, a second matching degree and a third matching degree of a fault feature library according to a matching algorithm;
and carrying out weighted calculation on the first matching degree, the second matching degree and the third matching degree to obtain a feature matching degree value, if the feature matching degree value is larger than a preset matching degree threshold value, matching the extraction result with a fault feature library, and if the feature matching degree value is not larger than the preset matching degree threshold value, not matching the extraction result with the fault feature library.
In some embodiments of the present application, obtaining the application evaluation degree of the preset neural model according to the test recognition result includes:
obtaining a first sample number A1 of correct recognition, a second sample number A2 of correct recognition of a fault image, a third number A3 of correct recognition of a normal image and a fourth number A4 of incorrect recognition of the normal image according to the test recognition result;
the calculation formula of the application evaluation degree is as follows:
wherein K is the application evaluation degree, and W is the total sample number of the test set identification result.
In some embodiments of the present application, calculating the confidence of the recognition result includes:
calculating sample evaluation degree according to a sample dimension value corresponding to the current recognition result, and determining the confidence coefficient of the recognition result according to the sample evaluation degree and the application evaluation degree;
the sample dimension value is calculated according to the sample resolution and the sample effective value;
the calculation formula of the sample evaluation degree is as follows:
wherein Q is the sample evaluation degree, L is the camera shooting range, V is the camera precision, beta 1 is the weight value corresponding to the sample resolution, r0 is the sample width and height, r1 is the width and height of the convolution kernel, S is the step value, p is the number of samples to be filled, and beta 2 is the weight value corresponding to the sample effective value;
the confidence coefficient is calculated according to the following formula:
T＝K*α1+Q*α2；
wherein, T is confidence, alpha 1 is the weight corresponding to the application evaluation, and alpha 2 is the weight corresponding to the sample evaluation.
In some embodiments of the present application, preprocessing a video image includes:
and carrying out gray processing on the video image to obtain a gray image, and removing noise data from the gray image to obtain an image to be detected, wherein the gray processing carries out weighted average processing on a plurality of colors of the video image and converts the colors into gray values.
In some embodiments of the present application, determining whether the identification result is in a fault feature library, and if so, sending an alarm signal includes:
and matching the identification result with the fault feature library to obtain a plurality of matching values, carrying out weighted calculation on the plurality of matching values to obtain a final matching degree value, if the final matching degree value is larger than a preset matching degree threshold value, matching the identification result with the fault feature library, wherein the identification result is in the fault feature library, sending an alarm signal, and if the final matching degree value is not larger than the preset matching degree threshold value, not matching the identification result with the fault feature library, and if the final matching degree value is not larger than the preset matching degree threshold value, not sending the alarm signal.
In some embodiments of the present application, a method for intelligent video monitoring of a coal conveying trestle is further included:
collecting video images of a monitoring area in real time, and preprocessing the video images to obtain images to be detected;
inputting the image to be detected into a trained neural network model, and carrying out feature extraction and recognition on the image to be detected by the neural network model to obtain a recognition result;
and calculating the confidence coefficient of the identification result, judging whether the identification result is in a fault feature library or not if the confidence coefficient is not smaller than a preset confidence coefficient threshold value, and sending an alarm signal if the identification result is in the fault feature library.
The invention provides a video intelligent monitoring system and a video intelligent monitoring method for a coal conveying trestle, which have the following beneficial effects compared with the prior art:
the invention discloses a video intelligent monitoring system and a video intelligent monitoring method for a coal conveying trestle, which are characterized in that a to-be-detected image is obtained by preprocessing a video image, the to-be-detected image is input into a trained neural network model, the neural network model performs feature extraction and recognition on the to-be-detected image to obtain a recognition result, the confidence coefficient of the recognition result is calculated, if the confidence coefficient is not smaller than a preset confidence coefficient threshold value, whether the recognition result is in a fault feature library is judged, if the confidence coefficient is in the fault feature library, an alarm signal is sent, if the confidence coefficient is smaller than the preset confidence coefficient threshold value, the to-be-detected image corresponding to the recognition result is input into the trained neural network model again until the confidence coefficient is not smaller than the preset confidence coefficient threshold value, and the problem that the recognition accuracy of a coal conveying image sample by the neural network model is low, the recognition result is not confirmed, and an error alarm signal is sent is solved.
Drawings
Fig. 1 shows a schematic structural diagram of a video intelligent monitoring system of a coal conveying trestle in an embodiment of the invention;
fig. 2 shows a flow diagram of a video intelligent monitoring method of a coal conveying trestle in an embodiment of the invention.
Detailed Description
The following describes in further detail the embodiments of the present invention with reference to the drawings and examples. The following examples are illustrative of the invention and are not intended to limit the scope of the invention.
In the description of the present application, it should be understood that the terms "center," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like indicate orientations or positional relationships based on the orientation or positional relationships shown in the drawings, merely to facilitate description of the present application and simplify the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present application.
The terms "first," "second," and the like, are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature. In the description of the present application, unless otherwise indicated, the meaning of "a plurality" is two or more.
In the description of the present application, it should be noted that, unless explicitly specified and limited otherwise, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be either fixedly connected, detachably connected, or integrally connected, for example; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the terms in this application will be understood by those of ordinary skill in the art in a specific context.
The following is a description of preferred embodiments of the invention, taken in conjunction with the accompanying drawings.
As shown in fig. 1, an embodiment of the present invention discloses a video intelligent monitoring system for a coal conveying trestle, the system comprising:
the acquisition module is used for acquiring video images of the monitoring area in real time, and preprocessing the video images to obtain images to be detected;
the identification module is used for inputting the image to be detected into a trained neural network model, and the neural network model performs feature extraction and identification on the image to be detected to obtain an identification result;
the judging module is used for calculating the confidence coefficient of the identification result, judging whether the identification result is in a fault feature library or not if the confidence coefficient is not smaller than a preset confidence coefficient threshold value, and sending an alarm signal if the identification result is in the fault feature library.
In this embodiment, preprocessing a video image to obtain a to-be-detected image, improving the sample resolution of the video image, performing feature extraction and recognition on the to-be-detected image by a trained neural network model, calculating the confidence coefficient of the recognition result, setting a preset confidence coefficient threshold according to historical data, confirming the recognition result in the historical data, wherein the preset confidence coefficient threshold is that the correct proportion of the recognition result reaches 80%, matching the extracted feature with a fault feature library when the confidence coefficient is not less than the preset confidence coefficient, judging whether the to-be-detected image corresponding to the extracted feature is matched with the fault feature in the fault feature library, and if so, sending an alarm signal, wherein the alarm signal comprises a fault type and a fault point, the fault point is determined on the basis of the to-be-detected image in the trained neural network model, thereby greatly improving the accuracy of the to-be-detected image and the neural network model, solving the problems that the neural network model has low recognition accuracy of the sample of the coal-conveying image and does not confirm the recognition result, and sending an error alarm signal is caused
In some embodiments of the present application, the trained neural network model comprises:
acquiring a historical video image, preprocessing the historical video image to obtain a historical detection image, and dividing the historical detection image into a training set and a test set, wherein the training set comprises a coal conveying stack bridge fault image and the test set comprises a coal conveying stack bridge fault image and a coal conveying stack bridge normal image;
constructing a preset neural network model, training the preset neural network model according to a training set, and extracting and identifying features of the training set by the preset neural model to obtain a first neural network model;
inputting the test set into a first neural network model to obtain a test identification result, and obtaining the application evaluation degree of the first neural model according to the test identification result;
when the application evaluation degree is not smaller than a preset application evaluation degree threshold, setting the first neural model as a trained neural network model, and when the application evaluation degree is smaller than the preset application evaluation degree threshold, adding the training set, and carrying out iterative training on the first neural network model.
In this embodiment, when feature extraction is performed on the training set, features of different scales are extracted in a resampling manner on the same image to be detected, feature integration is performed on features of different scales, and feature integration is mainly performed by integrating low-level physical features and high-level semantics, so that detection accuracy and robustness of the model are greatly improved.
In some embodiments of the present application, the feature extraction and identification of the training set by the preset neural model includes:
the preset neural model segments the historical video image in the training set, a target area is determined, the target area is extracted, an extraction result is obtained, the extraction result is matched with a fault feature library, if the extraction result is matched, the historical video image corresponding to the target area is a fault image, if the identification result is not matched, the historical video image corresponding to the target area is a normal image, and the target area is a suspected fault area;
the extraction result of the target area is as follows:
wherein,e is a preset neural model bias value, y is an activation function, and +.>For presetting a nerve model convolution kernel, h0 is a weight value of the convolution kernel, G i And (3) presetting a history video image input by a jth convolution layer in the nerve model.
In some embodiments of the present application, matching the extraction result with a fault signature library includes:
performing feature analysis on the extraction result to obtain a first preset feature set, a second preset feature set and a third preset feature set;
sequentially calculating a first preset feature set, a second preset feature set, a third preset feature set and a first matching degree, a second matching degree and a third matching degree of a fault feature library according to a matching algorithm;
and carrying out weighted calculation on the first matching degree, the second matching degree and the third matching degree to obtain a feature matching degree value, if the feature matching degree value is larger than a preset matching degree threshold value, matching the extraction result with a fault feature library, and if the feature matching degree value is not larger than the preset matching degree threshold value, not matching the extraction result with the fault feature library.
In this embodiment, the importance degrees of the feature sets are a first preset feature set, a second preset feature set, and a third preset feature set in sequence, the weight value of the first matching degree is greater than the weight value of the second matching degree, the weight value of the second matching degree is greater than the weight value of the third matching degree, and the preset matching degree threshold is set according to the matching degree matched with the fault feature library in the historical data.
In some embodiments of the present application, obtaining the application evaluation degree of the preset neural model according to the test recognition result includes:
obtaining a first sample number A1 of correct recognition, a second sample number A2 of correct recognition of a fault image, a third number A3 of correct recognition of a normal image and a fourth number A4 of incorrect recognition of the normal image according to the test recognition result;
the calculation formula of the application evaluation degree is as follows:
wherein K is the application evaluation degree, and W is the total sample number of the test set identification result.
In some embodiments of the present application, calculating the confidence of the recognition result includes:
calculating sample evaluation degree according to a sample dimension value corresponding to the current recognition result, and determining the confidence coefficient of the recognition result according to the sample evaluation degree and the application evaluation degree;
the sample dimension value is calculated according to the sample resolution and the sample effective value;
the calculation formula of the sample evaluation degree is as follows:
wherein Q is the sample evaluation degree, L is the camera shooting range, V is the camera precision, beta 1 is the weight value corresponding to the sample resolution, r0 is the sample width and height, r1 is the width and height of the convolution kernel, S is the step value, p is the number of samples to be filled, and beta 2 is the weight value corresponding to the sample effective value;
the confidence coefficient is calculated according to the following formula:
T＝K*α1+Q*α2；
wherein, T is confidence, alpha 1 is the weight corresponding to the application evaluation, and alpha 2 is the weight corresponding to the sample evaluation.
In some embodiments of the present application, preprocessing a video image includes:
and carrying out gray processing on the video image to obtain a gray image, and removing noise data from the gray image to obtain an image to be detected, wherein the gray processing carries out weighted average processing on a plurality of colors of the video image and converts the colors into gray values.
In some embodiments of the present application, determining whether the identification result is in a fault feature library, and if so, sending an alarm signal includes:
and matching the identification result with the fault feature library to obtain a plurality of matching values, carrying out weighted calculation on the plurality of matching values to obtain a final matching degree value, if the final matching degree value is larger than a preset matching degree threshold value, matching the identification result with the fault feature library, wherein the identification result is in the fault feature library, sending an alarm signal, and if the final matching degree value is not larger than the preset matching degree threshold value, not matching the identification result with the fault feature library, and if the final matching degree value is not larger than the preset matching degree threshold value, not sending the alarm signal.
In some embodiments of the present application, a method for intelligent video monitoring of a coal conveying trestle is further included:
collecting video images of a monitoring area in real time, and preprocessing the video images to obtain images to be detected;
inputting the image to be detected into a trained neural network model, and carrying out feature extraction and recognition on the image to be detected by the neural network model to obtain a recognition result;
and calculating the confidence coefficient of the identification result, judging whether the identification result is in a fault feature library or not if the confidence coefficient is not smaller than a preset confidence coefficient threshold value, and sending an alarm signal if the identification result is in the fault feature library.
In summary, the invention discloses a video intelligent monitoring system of a coal conveying trestle, which comprises: the acquisition module is used for acquiring video images of the monitoring area in real time, and preprocessing the video images to obtain images to be detected; the identification module is used for inputting the image to be detected into a trained neural network model, and the neural network model performs feature extraction and identification on the image to be detected to obtain an identification result; the judging module is used for calculating the confidence coefficient of the identification result, judging whether the identification result is in a fault feature library or not if the confidence coefficient is not smaller than a preset confidence coefficient threshold value, and sending an alarm signal if the identification result is in the fault feature library.
According to the method, the image to be detected is obtained by preprocessing the video image, the image to be detected is input into the trained neural network model, the neural network model performs feature extraction and recognition on the image to be detected to obtain a recognition result, the confidence coefficient of the recognition result is calculated, whether the recognition result is in a fault feature library is judged again if the confidence coefficient is not smaller than a preset confidence coefficient threshold value, an alarm signal is sent if the confidence coefficient is in the fault feature library, the image to be detected corresponding to the recognition result is input into the trained neural network model again until the confidence coefficient is not smaller than the preset confidence coefficient threshold value, and the problem that the neural network model has low recognition accuracy of a coal conveying image sample and does not confirm the recognition result and sends an error alarm signal is solved.
In the description of the above embodiments, particular features, structures, materials, or characteristics may be combined in any suitable manner in any one or more embodiments or examples.
Although the invention has been described hereinabove with reference to embodiments, various modifications thereof may be made and equivalents may be substituted for elements thereof without departing from the scope of the invention. In particular, the features of the disclosed embodiments may be combined with each other in any manner as long as there is no structural conflict, and the entire description of these combinations is not made in the present specification merely for the sake of omitting the descriptions and saving resources. Therefore, it is intended that the invention not be limited to the particular embodiment disclosed, but that the invention will include all embodiments falling within the scope of the appended claims.
Those of ordinary skill in the art will appreciate that: the above is only a preferred embodiment of the present invention, and the present invention is not limited thereto, but it is to be understood that the present invention is described in detail with reference to the foregoing embodiments, and modifications and equivalents of some of the technical features described in the foregoing embodiments may be made by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (9)

1. A video intelligent monitoring system of coal conveying trestle, which is characterized by comprising:
the acquisition module is used for acquiring video images of the monitoring area in real time, and preprocessing the video images to obtain images to be detected;
the identification module is used for inputting the image to be detected into a trained neural network model, and the neural network model performs feature extraction and identification on the image to be detected to obtain an identification result;
the judging module is used for calculating the confidence coefficient of the identification result, judging whether the identification result is in a fault feature library or not if the confidence coefficient is not smaller than a preset confidence coefficient threshold value, and sending an alarm signal if the identification result is in the fault feature library.
2. The video intelligent monitoring system of a coal delivery trestle according to claim 1, characterized in that the trained neural network model comprises:
acquiring a historical video image, preprocessing the historical video image to obtain a historical detection image, and dividing the historical detection image into a training set and a test set, wherein the training set comprises a coal conveying stack bridge fault image and the test set comprises a coal conveying stack bridge fault image and a coal conveying stack bridge normal image;
constructing a preset neural network model, training the preset neural network model according to a training set, and extracting and identifying features of the training set by the preset neural model to obtain a first neural network model;
inputting the test set into a first neural network model to obtain a test identification result, and obtaining the application evaluation degree of the first neural model according to the test identification result;
when the application evaluation degree is not smaller than a preset application evaluation degree threshold, setting the first neural model as a trained neural network model, and when the application evaluation degree is smaller than the preset application evaluation degree threshold, adding the training set, and carrying out iterative training on the first neural network model.
3. The intelligent video monitoring system of a coal conveying trestle according to claim 2, characterized in that said preset neural model performs feature extraction and recognition on said training set, and comprises:
the preset neural model segments the historical video image in the training set, a target area is determined, the target area is extracted, an extraction result is obtained, the extraction result is matched with a fault feature library, if the extraction result is matched, the historical video image corresponding to the target area is a fault image, if the identification result is not matched, the historical video image corresponding to the target area is a normal image, and the target area is a suspected fault area;
the extraction result of the target area is as follows:
wherein P is i j For the extraction result of the target area output by the ith channel of the jth convolution layer in the preset nerve model, E is the preset nerve model bias value, y is the activation function,for presetting a nerve model convolution kernel, h0 is a weight value of the convolution kernel, G i And (3) presetting a history video image input by a jth convolution layer in the nerve model.
4. The intelligent video monitoring system of a coal conveying trestle according to claim 3, characterized in that matching the extraction result with a fault feature library comprises:
performing feature analysis on the extraction result to obtain a first preset feature set, a second preset feature set and a third preset feature set;
sequentially calculating a first preset feature set, a second preset feature set, a third preset feature set and a first matching degree, a second matching degree and a third matching degree of a fault feature library according to a matching algorithm;
and carrying out weighted calculation on the first matching degree, the second matching degree and the third matching degree to obtain a feature matching degree value, if the feature matching degree value is larger than a preset matching degree threshold value, matching the extraction result with a fault feature library, and if the feature matching degree value is not larger than the preset matching degree threshold value, not matching the extraction result with the fault feature library.
5. The intelligent video monitoring system of a coal conveying trestle according to claim 3, wherein the obtaining the application evaluation degree of the preset neural model according to the test recognition result comprises:
obtaining a first sample number A1 of correct recognition, a second sample number A2 of correct recognition of a fault image, a third number A3 of correct recognition of a normal image and a fourth number A4 of incorrect recognition of the normal image according to the test recognition result;
the calculation formula of the application evaluation degree is as follows:
wherein K is the application evaluation degree, and W is the total sample number of the test set identification result.
6. The video intelligent monitoring system of a coal delivery trestle according to claim 5, wherein calculating the confidence of the recognition result comprises:
calculating sample evaluation degree according to a sample dimension value corresponding to the current recognition result, and determining the confidence coefficient of the recognition result according to the sample evaluation degree and the application evaluation degree;
the sample dimension value is calculated according to the sample resolution and the sample effective value;
the calculation formula of the sample evaluation degree is as follows:
wherein Q is the sample evaluation degree, L is the camera shooting range, V is the camera precision, beta 1 is the weight value corresponding to the sample resolution, r0 is the sample width and height, r1 is the width and height of the convolution kernel, S is the step value, p is the number of samples to be filled, and beta 2 is the weight value corresponding to the sample effective value;
the confidence coefficient is calculated according to the following formula:
T＝K*α1+Q*α2；
wherein, T is confidence, alpha 1 is the weight corresponding to the application evaluation, and alpha 2 is the weight corresponding to the sample evaluation.
7. The intelligent video monitoring system of a coal conveying trestle according to claim 6, characterized in that preprocessing the video image comprises:
and carrying out gray processing on the video image to obtain a gray image, and removing noise data from the gray image to obtain an image to be detected, wherein the gray processing carries out weighted average processing on a plurality of colors of the video image and converts the colors into gray values.
8. The intelligent video monitoring method of a coal conveying trestle according to claim 4, wherein determining whether the identification result is in a fault feature library, and if so, sending an alarm signal comprises:
and matching the identification result with the fault feature library to obtain a plurality of matching values, carrying out weighted calculation on the plurality of matching values to obtain a final matching degree value, if the final matching degree value is larger than a preset matching degree threshold value, matching the identification result with the fault feature library, wherein the identification result is in the fault feature library, sending an alarm signal, and if the final matching degree value is not larger than the preset matching degree threshold value, not matching the identification result with the fault feature library, and if the final matching degree value is not larger than the preset matching degree threshold value, not sending the alarm signal.
9. The intelligent video monitoring method for the coal conveying trestle is characterized by comprising the following steps of:
collecting video images of a monitoring area in real time, and preprocessing the video images to obtain images to be detected;
inputting the image to be detected into a trained neural network model, and carrying out feature extraction and recognition on the image to be detected by the neural network model to obtain a recognition result;
and calculating the confidence coefficient of the identification result, judging whether the identification result is in a fault feature library or not if the confidence coefficient is not smaller than a preset confidence coefficient threshold value, and sending an alarm signal if the identification result is in the fault feature library.
CN202310967424.4A 2023-08-02 2023-08-02 Video intelligent monitoring system and method for coal conveying trestle Pending CN117392594A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310967424.4A CN117392594A (en) 2023-08-02 2023-08-02 Video intelligent monitoring system and method for coal conveying trestle

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310967424.4A CN117392594A (en) 2023-08-02 2023-08-02 Video intelligent monitoring system and method for coal conveying trestle

Publications (1)

Publication Number Publication Date
CN117392594A true CN117392594A (en) 2024-01-12

Family

ID=89463758

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310967424.4A Pending CN117392594A (en) 2023-08-02 2023-08-02 Video intelligent monitoring system and method for coal conveying trestle

Country Status (1)

Country Link
CN (1) CN117392594A (en)

Similar Documents

Publication Publication Date Title
CN112818988B (en) Automatic identification reading method and system for pointer instrument
CN109948469A (en) The automatic detection recognition method of crusing robot instrument based on deep learning
CN113989257A (en) Electric power comprehensive pipe gallery settlement crack identification method based on artificial intelligence technology
CN113658192A (en) Multi-target pedestrian track acquisition method, system, device and medium
CN110852164A (en) YOLOv 3-based method and system for automatically detecting illegal building
CN111523416A (en) Vehicle early warning method and device based on highway ETC portal
CN116612098B (en) Insulator RTV spraying quality evaluation method and device based on image processing
CN113688817A (en) Instrument identification method and system for automatic inspection
CN112132157A (en) Raspberry pie-based gait face fusion recognition method
CN109447946B (en) Overhead communication optical cable abnormality detection method
CN117392594A (en) Video intelligent monitoring system and method for coal conveying trestle
CN116385866B (en) SAR image-based railway line color steel house change detection method and device
CN111539931A (en) Appearance abnormity detection method based on convolutional neural network and boundary limit optimization
CN114581407B (en) Self-adaptive defect detection method for photovoltaic module
CN114061476B (en) Method for detecting deflection of insulator of power transmission line
CN113747270B (en) Cloud data center thing allies oneself with management and control system
CN112508946B (en) Cable tunnel anomaly detection method based on antagonistic neural network
CN112257567B (en) Training of behavior recognition network, behavior recognition method and related equipment
CN114743048A (en) Method and device for detecting abnormal straw picture
CN113850173A (en) Method, device and medium for eliminating repeated positioning of faulty photovoltaic module
CN112200036A (en) Student behavior remote monitoring method and system
CN113155293B (en) Human body remote sensing temperature measurement monitoring and recognition system based on unmanned aerial vehicle
CN115546793B (en) Automatic reading method and system for water gauge scales and electronic equipment
CN116805435B (en) Intelligent inspection device for motor room
CN109726724B (en) Water gauge image feature weighted learning identification method under shielding condition

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