CN113469937A - Pipe gallery abnormal position positioning method and system based on pipe gallery video and two-dimensional code detection - Google Patents
Pipe gallery abnormal position positioning method and system based on pipe gallery video and two-dimensional code detection Download PDFInfo
- Publication number
- CN113469937A CN113469937A CN202110572994.4A CN202110572994A CN113469937A CN 113469937 A CN113469937 A CN 113469937A CN 202110572994 A CN202110572994 A CN 202110572994A CN 113469937 A CN113469937 A CN 113469937A
- Authority
- CN
- China
- Prior art keywords
- dimensional code
- pipe gallery
- abnormal
- picture
- model
- 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
Links
Images
Classifications
-
- 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
- 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
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06K—GRAPHICAL DATA READING; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
- G06K7/00—Methods or arrangements for sensing record carriers, e.g. for reading patterns
- G06K7/10—Methods or arrangements for sensing record carriers, e.g. for reading patterns by electromagnetic radiation, e.g. optical sensing; by corpuscular radiation
- G06K7/14—Methods or arrangements for sensing record carriers, e.g. for reading patterns by electromagnetic radiation, e.g. optical sensing; by corpuscular radiation using light without selection of wavelength, e.g. sensing reflected white light
- G06K7/1404—Methods for optical code recognition
- G06K7/1408—Methods for optical code recognition the method being specifically adapted for the type of code
- G06K7/1417—2D bar codes
-
- 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
- G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Toxicology (AREA)
- Electromagnetism (AREA)
- Quality & Reliability (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method and a system for positioning abnormal positions of a pipe gallery based on pipe gallery video and two-dimensional code detection, wherein the method comprises the following steps: inputting a picture of the pipe gallery to be positioned and identified into a pre-trained pipe gallery abnormity classification neural network model, and obtaining a pipe gallery abnormity classification result; if the pipe gallery abnormity classification result is abnormal, inputting the picture of the pipe gallery to be positioned and identified into a pre-trained two-dimensional code detection model, and acquiring the position of the two-dimensional code in the picture of the pipe gallery to be positioned and identified; acquiring and identifying a two-dimensional code picture based on the position of the two-dimensional code to acquire two-dimensional code pre-recorded information; based on the two-dimensional code is recorded information in advance and is accomplished piping lane abnormal position location. According to the technical scheme provided by the invention, the detection of the abnormal position of the pipe gallery can be completed based on the two-dimensional code detection and the collection of the pipe gallery video; compared with the traditional positioning mode at present, the method has higher safety, reliability and accuracy.
Description
Technical Field
The invention belongs to the technical field of computer digital image processing and pattern recognition, and particularly relates to a method and a system for positioning an abnormal position of a pipe gallery based on pipe gallery video and two-dimensional code detection.
Background
The utility tunnel is that the pipelines such as water, electricity, gas and heating letter are unified to be inserted into the pipe gallery to be favorable to operation, management, maintenance and inquiry. The pipeline laid in the utility tunnel is easy to lose due to the complex pipe gallery environment, and the phenomena of disordered overlapping of the pipeline, falling of a pipeline support and the like exist in common abnormal conditions.
At present, the location to the abnormal position in the utility tunnel can be through two kinds of methods, one is that the abnormal position is reported to the mode through manual record, one is that the GPS through the piping lane robot fixes a position, and the defect that these two kinds of traditional modes exist includes:
(1) for the mode of manually recording the abnormal position, the environment of the pipe gallery is dangerous, the manual recording is difficult to guarantee the safety of personnel, and the time and labor consumption are low;
(2) for the method of positioning the abnormal position by using the GPS, the pipe gallery is generally positioned at the ground bottom of a city, the GPS signal in the pipe gallery is very weak, the abnormal position is difficult to position, and the reliability is poor; in addition, the error of the GPS is large, and the accuracy performance of abnormal position positioning is poor.
Disclosure of Invention
The invention aims to provide a method and a system for positioning an abnormal position of a pipe gallery based on pipe gallery video and two-dimensional code detection, so as to solve one or more technical problems. According to the technical scheme provided by the invention, the detection of the abnormal position of the pipe gallery can be completed based on the two-dimensional code detection and the collection of the pipe gallery video; compared with the traditional positioning mode at present, the method has higher safety, reliability and accuracy.
In order to achieve the purpose, the invention adopts the following technical scheme:
the invention discloses a pipe gallery abnormal position positioning method based on pipe gallery video and two-dimensional code detection, which comprises the following steps:
inputting a picture of the pipe gallery to be positioned and identified into a pre-trained pipe gallery abnormity classification neural network model, and obtaining a pipe gallery abnormity classification result;
if the pipe gallery abnormity classification result is abnormal, inputting the picture of the pipe gallery to be positioned and identified into a pre-trained two-dimensional code detection model, and acquiring the position of the two-dimensional code in the picture of the pipe gallery to be positioned and identified;
acquiring and identifying a two-dimensional code picture based on the position of the two-dimensional code to acquire two-dimensional code pre-recorded information; based on the two-dimensional code is recorded information in advance and is accomplished piping lane abnormal position location.
The invention is further improved in that the step of obtaining the picture of the pipe gallery to be positioned and identified comprises the following steps: pipe gallery video based on the visual camera collection that sets up on the pipe gallery robot acquires the identification pipe gallery picture of waiting to fix a position.
The invention further improves that the pipe gallery abnormal classification neural network model is Resnet, Densenet, VGG or Alexnet.
The invention is further improved in that the two-dimensional code detection model is YOLO, Fast-RCNN or SSD.
The invention has the further improvement that the step of obtaining the trained neural network model for classifying the abnormality of the pipe gallery comprises the following steps:
acquiring a tube corridor abnormal learning sample set;
training the pipe gallery abnormal classification neural network model based on the pipe gallery abnormal learning sample set to obtain a trained pipe gallery abnormal classification neural network model;
and performing pruning search on the trained pipe gallery abnormal classification neural network model by using a netadapt method, converting the model into an ONNX format after the search is finished, and quantizing to obtain an RKNN model and obtain the trained pipe gallery abnormal classification neural network model.
The invention has the further improvement that the acquisition step of the trained two-dimensional code detection model comprises the following steps:
acquiring a two-dimensional code detection learning sample set;
training the two-dimensional code detection model based on the two-dimensional code detection learning sample set to obtain a trained two-dimensional code detection model;
and performing pruning search on the trained two-dimensional code detection model by using a netadapt method, converting the model into an ONNX format after the search is finished, and quantizing to obtain an RKNN model and obtain the trained two-dimensional code detection model.
The invention is further improved in that if the result of the abnormal classification of the pipe gallery is abnormal, the method further comprises the following steps:
and judging whether the abnormal value exceeds a preset threshold value, and if so, outputting the to-be-positioned identification pipe gallery picture and information pre-recorded by the two-dimensional code on the picture and giving an alarm.
The further improvement of the invention is that the steps of obtaining and identifying the two-dimension code picture based on the position of the two-dimension code and obtaining the pre-recorded information of the two-dimension code specifically comprise:
acquiring a two-dimensional code picture on a to-be-positioned identification pipe gallery picture based on the position of the two-dimensional code;
after the definition of the two-dimensional code picture is enhanced through a hyper-resolution network based on deep learning, the number information pre-recorded in the two-dimensional code is identified by using pyzbar, and the number information is used for representing the information of the preset position in the pipe gallery.
In the super-division network, a close connection mode is used for connecting the shallow features and the deep features of the network; and setting a residual error structure for accelerating network convergence.
The invention discloses a pipe gallery abnormal position positioning system based on pipe gallery video and two-dimensional code detection, which comprises:
the pipe gallery abnormity classification result acquisition module is used for inputting the picture of the pipe gallery to be positioned and identified into a pre-trained pipe gallery abnormity classification neural network model to acquire a pipe gallery abnormity classification result;
the two-dimensional code position acquisition module is used for inputting the picture of the pipe gallery to be positioned and identified into a pre-trained two-dimensional code detection model when the abnormal classification result of the pipe gallery is abnormal, and acquiring the position of the two-dimensional code in the picture of the pipe gallery to be positioned and identified;
the identification positioning module is used for acquiring and identifying a two-dimensional code picture according to the position of the two-dimensional code to acquire two-dimensional code pre-recorded information; based on the two-dimensional code is recorded information in advance and is accomplished piping lane abnormal position location.
Compared with the prior art, the invention has the following beneficial effects:
in the invention, a deep learning network model is generated based on sample learning, and the abnormal position of the pipe gallery is determined through a pipe gallery abnormal classification network model and a two-dimensional code detection and identification model. The method for positioning the abnormal position of the pipe gallery based on the video of the pipe gallery (for example, the abnormal position is obtained by automatic inspection through a robot camera) is safer, more reliable and more accurate compared with the current method of manual inspection. In addition, compared with a method using a GPS (global positioning system) for positioning, the position interval of the two-dimensional code picture preset in the pipe gallery can be set more densely (for example, set according to the preset precision requirement), the positioning error is smaller, the precision requirement can be met, and the condition that the GPS is disconnected can not be generated (once the model is trained, the position where the abnormality occurs can be recorded in an offline mode, and the abnormal position can also be stored by a network uploading server).
In the invention, pictures are obtained based on a pipe gallery video obtained by a visual camera, and are made into detection and analysis samples, a deep learning network model is generated based on sample learning, the trained network model is quantized and pruned, and intelligent front-end equipment is used for detecting the two-dimensional codes and pipe gallery abnormity (such as electric abnormity detection). According to the invention, the two-dimensional code is detected to obtain the position of the robot (provided with the visual camera for collecting the video of the pipe gallery), so that the robot has higher safety and reliability compared with manual maintenance; in addition, the method also has higher instantaneity (for example, dangerous information is uploaded immediately by reporting the background of the server).
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art are briefly introduced below; it is obvious that the drawings in the following description are some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.
Fig. 1 is a schematic flow chart of a pipe gallery abnormal position positioning method based on pipe gallery video and two-dimensional code detection according to an embodiment of the present invention;
fig. 2 is a schematic diagram of two-dimensional code detection in the embodiment of the present invention.
Detailed Description
In order to make the purpose, technical effect and technical solution of the embodiments of the present invention clearer, the following clearly and completely describes the technical solution of the embodiments of the present invention with reference to the drawings in the embodiments of the present invention; it is to be understood that the described embodiments are only some of the embodiments of the present invention. Other embodiments, which can be derived by one of ordinary skill in the art from the disclosed embodiments without inventive faculty, are intended to be within the scope of the invention.
Referring to fig. 1, a method for positioning an abnormal position of a pipe rack based on pipe rack video and two-dimensional code detection according to an embodiment of the present invention includes the following steps:
step 1: acquiring a two-dimensional code detection and pipe gallery abnormal learning sample set; illustratively, learning samples of two-dimensional code detection and pipe gallery abnormity are obtained according to a visual camera arranged on a pipe gallery robot, and the learning samples are used for training a classification depth neural network model of the two-dimensional code detection and the pipe gallery abnormity; a preset two-dimensional code picture (which can be set according to actual demand precision or preset intervals) is arranged at a preset position in the pipe gallery, information pre-recorded in the two-dimensional code picture is a preset number, and the preset number corresponds to the position of the two-dimensional code picture in the pipe gallery;
step 2: training a detection model for detecting the two-dimensional code; the neural network adopted by the detection model of the embodiment of the invention is YOLO, Fast-RCNN or SSD; optionally, storing the trained network model to a local hard disk;
and step 3: constructing a deep neural network, and training the neural network by using data of the pipe gallery abnormal learning sample set; preferably, the neural network adopted by the embodiment of the invention comprises Resnet, Densenet, VGG or Alexnet, and the trained deep network model is stored on the local hard disk;
and 4, step 4: pruning and searching the trained network model by using a netadapt method, converting the model into an ONNX format after the searching is finished, and quantizing to obtain an RKNN model;
and 5: directly acquiring images from a camera or a video file on a local hard disk, and performing two-dimensional code detection and pipe gallery anomaly detection on a newly acquired image to be identified by using a trained network model;
detecting whether the pipe gallery is abnormal or not through an abnormality detection model, and detecting the position of the two-dimensional code through a two-dimensional code detection model;
if the pipe gallery is abnormal, the serial number detected by the two-dimensional code is returned at the moment, and the current image of the pipe gallery robot is stored and uploaded to the background server.
In step 4 of the embodiment of the present invention, the step of converting the trained network model into an ONNX format and then quantizing the ONNX format to obtain the RKNN model specifically includes:
(1) the trained model is converted into ONNX format, and the model parameters are quantized to int8 type.
(2) And converting the quantized model into an RKNN format through an RKNNtoolkit so as to operate in the embedded front end.
In step 5 of the embodiment of the invention, the embedded device detects that the two-dimensional code inputs a complete camera picture through the quantized model obtained in step 4, and classifies the left and right images of the pipe gallery which are input abnormally.
In the embodiment of the invention, after the two-dimensional code is obtained, the definition of the two-dimensional code is enhanced through a super-resolution technology based on deep learning, and then the pyzbar is used for identifying the condition of the number in the two-dimensional code, wherein the number of the two-dimensional code represents the position of the robot, and whether the left side and the right side of the pipe gallery are abnormal or not is judged in real time.
The hyper-division network of the embodiment of the invention comprises FSRCNN, ESPCN, VDSR, EDSR and IDN, and the algorithm needs to be improved because the hyper-division network needs to be deployed on embedded equipment which has limited computing power and cannot carry out real-time reasoning on the algorithm. Firstly, the depth and the width of the network need to be compressed, but the precision loss is brought, and in order to reduce the loss, the invention adopts a closely connected structure in densenet and a network structure for residual learning in resnet to improve the compressed network. And the network shallow feature is connected with the deep feature by using a close connection mode, so that the small target two-dimensional code in the picture can be identified. By using the residual error structure, the network convergence speed can be increased, and the performance of the hyper-division network can be improved.
In the embodiment of the invention, the image acquired by the camera of the pipe gallery robot is classified abnormally through the model in the step 3, if the image is in an abnormal state, the position of the two-dimensional code is obtained through the detection model obtained in the step 2, and finally the position information contained in the two-dimensional code is identified through the position of the two-dimensional code and reported to the server.
In the embodiment of the invention, when the abnormal threshold value of the pipe gallery is higher than the set threshold value, the image of the pipe gallery at the moment and the number of the two-dimensional code are sent to the background.
The method according to the above embodiment of the present invention includes: training and detecting a two-dimensional code and a pipe gallery abnormity classification model; searching and quantifying the detection model and the abnormal classification, and deploying the detection model and the abnormal classification on an intelligent front end; the two-dimensional code is clearer through super-resolution counting based on deep learning; identifying the two-dimensional code through pyzbar and extracting a number representing position information; and reporting to the background server according to the abnormal request condition of the pipe gallery. Compared with the traditional method, the method can improve the safety, reliability and accuracy of positioning identification.
Referring to fig. 1 and 2, a method for positioning an abnormal position of a pipe rack based on pipe rack video and two-dimensional code detection according to an embodiment of the present invention includes:
in step 1 of the embodiment of the present invention, the step of obtaining a two-dimensional code and a pipe gallery abnormal learning sample set specifically includes:
through installing the vision camera on the piping lane robot, can obtain each frame piping lane image (contained in the image at the predetermined two-dimensional code picture of piping lane preset), carry out attribute and position mark to it after obtaining the picture. The method is manufactured into a VOC or COCO data set, and the subsequent training task of the detection model is facilitated. And cut the both sides of piping lane, left side rectangle frame: (80, 150), (680, 1070), right matrix box: (1100,150), (1800, 950); as classification samples of normal and abnormal.
In step 2 of the embodiment of the present invention, the step of training the detection model for detecting the two-dimensional code specifically includes:
firstly, a detection model is selected, such as YOLO, SSD, fasternn, etc., and the depth model is trained by the sample obtained in step 1. Because the proportion of the two-dimensional code to the whole picture is smaller, a smaller anchor and a smaller IOU threshold value can be selected to improve the recall rate.
Because the proportion of the two-dimensional code relative to the whole picture is smaller, a smaller anchor can be used in the detection model in the step 2, and the IOU threshold of the positive and negative samples in the network is reduced to improve the recall rate of the two-dimensional code detection.
In step 3 of the embodiment of the present invention, the step of training the pipe gallery abnormal classification model specifically includes:
firstly, selecting a classification model, namely Resnet, Densenet, VGG, Alexnet and the like, training a depth model by using the classification sample obtained in the step 1, and solving the problem of abnormal and normal unbalance by adopting an oversampling method.
For the classified depth model, each class of training samples of the classification model can be oversampled until the number of each class is consistent, so that the problem of unbalanced number of abnormal samples and normal samples is solved, and the performance of the classification network is improved.
In step 4 of the embodiment of the present invention, pruning search is performed on the trained network model (referring to the detection model obtained in step 2 and the classification model obtained in step 3) by using a netadapt method, and after the search is completed, the model is converted into an ONNX format and then quantized, so as to obtain an RKNN model, including:
firstly, the detection model obtained by training in the step 2 and the step 3 is subjected to pruning search by a netadapt method, so that the parameter quantity of the model is further reduced. After the search, the model is converted to ONNX format and quantized, resulting in the RKNN model that can be run on the smart front-end.
In step 5 of the embodiment of the present invention, directly obtaining an image from a camera or from a video file on a local hard disk, and detecting the two-dimensional code position and the pipe rack abnormality by using a trained network model, includes:
will install the whole image of the camera on piping lane robot as the input that detects the two-dimensional code, will whole image divide into about two, left side rectangle frame: (80, 150), (680, 1070), right matrix box: (1100,150), (1800, 950) and feeding the pictures of the two positions as input into an anomaly classification network.
In step 6 of the embodiment of the present invention, whether the pipe gallery is abnormal is detected through the abnormality detection model, and the step of detecting the position of the two-dimensional code through the two-dimensional code detection model specifically includes:
after the position of the two-dimensional code is obtained in step 5, the network includes technologies such as Dense connection (Dense Concat), deep Convolution (deep Convolution), reverse Convolution (Deconvolution), Residual Learning (Residual Learning) to improve the performance of the model and identify the number of the representative position in the two-bit code, so that which position the pipe gallery robot specifically reaches can be determined through the hyper-division network based on deep Learning. And after the abnormal probability of the left and right pipe racks is obtained, if the abnormal probability exceeds a preset threshold value, judging that the part of the pipe rack is abnormal at the moment.
The detection model of the two-dimensional code can be modified through the SSD network structure, the picture that the pipe gallery robot camera was gathered is input to the detection model, what output is the position that detects the two-dimensional code, can read the positional information who contains in the two-dimensional code through pyzbr or based on the method of deep learning after detecting the position of two-dimensional code. Preferably, the speed of positioning the two-dimensional code is increased by reducing the characteristics of the original SSD model output.
In the embodiment of the present invention, if the pipe rack is abnormal, reporting the abnormal condition of the pipe rack to the background by combining the number information of the two-dimensional code, including: when the abnormal probability of the cable in the pipe gallery exceeds a certain threshold value, the position number of the obtained pipe gallery robot and the camera image of the robot at the moment are uploaded to a server at a background, and operation and maintenance personnel can conveniently inquire and maintain the position number and the camera image.
In summary, the present invention discloses a method for intelligently managing or positioning a pipe gallery video based on two-dimensional code positioning, which includes: carrying out lightweight design on the detection model; searching and quantifying a detection model, and deploying the detection model on an intelligent front end; and (4) obtaining the abnormal position of the pipe gallery by combining the two-dimensional code, and reporting the abnormal picture to a background. According to the technical scheme provided by the embodiment of the invention, the detection of the pipe gallery cable can be carried out through the intelligent front-end equipment which is easy to deploy, the detection speed is further accelerated through a quantification and searching method, the abnormal position of the pipe gallery is extracted by combining the two-dimensional code, and the abnormal picture is reported to the background, so that the query and maintenance of operation and maintenance personnel are facilitated.
As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art can make modifications and equivalents to the embodiments of the present invention without departing from the spirit and scope of the present invention, which is set forth in the claims of the present application.
Claims (10)
1. The utility model provides a pipe gallery abnormal position positioning method based on pipe gallery video and two-dimensional code detection which characterized in that includes following steps:
inputting a picture of the pipe gallery to be positioned and identified into a pre-trained pipe gallery abnormity classification neural network model, and obtaining a pipe gallery abnormity classification result;
if the pipe gallery abnormity classification result is abnormal, inputting the picture of the pipe gallery to be positioned and identified into a pre-trained two-dimensional code detection model, and acquiring the position of the two-dimensional code in the picture of the pipe gallery to be positioned and identified;
and acquiring and identifying a two-dimension code picture based on the position of the two-dimension code, acquiring pre-recorded information of the two-dimension code, and finishing positioning of the abnormal position of the pipe gallery based on the pre-recorded information of the two-dimension code.
2. The method for positioning the abnormal position of the pipe gallery based on the pipe gallery video and the two-dimensional code detection according to claim 1, wherein the step of obtaining the picture of the pipe gallery to be positioned and identified comprises the following steps: pipe gallery video based on the visual camera collection that sets up on the pipe gallery robot acquires the identification pipe gallery picture of waiting to fix a position.
3. The method according to claim 1, wherein the pipe rack anomaly location method based on pipe rack video and two-dimensional code detection is characterized in that the pipe rack anomaly classification neural network model is Resnet, Densenet, VGG or Alexnet.
4. The method as claimed in claim 1, wherein the two-dimensional code detection model is YOLO, Fast-RCNN or SSD.
5. The method for positioning the abnormal position of the pipe rack based on the pipe rack video and the two-dimensional code detection according to claim 1, wherein the step of obtaining the trained neural network model for the abnormal classification of the pipe rack comprises the following steps:
acquiring a tube corridor abnormal learning sample set;
training the pipe gallery abnormal classification neural network model based on the pipe gallery abnormal learning sample set to obtain a trained pipe gallery abnormal classification neural network model;
and performing pruning search on the trained pipe gallery abnormal classification neural network model by using a netadapt method, converting the model into an ONNX format after the search is finished, and quantizing to obtain an RKNN model and obtain the trained pipe gallery abnormal classification neural network model.
6. The method for positioning the abnormal position of the pipe rack based on the pipe rack video and the two-dimensional code detection according to claim 1, wherein the step of acquiring the trained two-dimensional code detection model comprises the following steps:
acquiring a two-dimensional code detection learning sample set;
training the two-dimensional code detection model based on the two-dimensional code detection learning sample set to obtain a trained two-dimensional code detection model;
and performing pruning search on the trained two-dimensional code detection model by using a netadapt method, converting the model into an ONNX format after the search is finished, and quantizing to obtain an RKNN model and obtain the trained two-dimensional code detection model.
7. The method for positioning the abnormal position of the pipe rack based on the pipe rack video and the two-dimensional code detection according to claim 1, wherein if the abnormal classification result of the pipe rack is abnormal, the method further comprises:
and judging whether the abnormal value exceeds a preset threshold value, and if so, outputting the to-be-positioned identification pipe gallery picture and information pre-recorded by the two-dimensional code on the picture and giving an alarm.
8. The method for positioning the abnormal position of the pipe rack based on the pipe rack video and the two-dimensional code detection according to claim 1, wherein the two-dimensional code picture is obtained and recognized based on the position of the two-dimensional code, and the step of obtaining the pre-recorded information of the two-dimensional code specifically comprises the following steps:
acquiring a two-dimensional code picture on a to-be-positioned identification pipe gallery picture based on the position of the two-dimensional code;
after the definition of the two-dimensional code picture is enhanced through a hyper-resolution network based on deep learning, the number information pre-recorded in the two-dimensional code is identified by using pyzbar, and the number information is used for representing the information of the preset position in the pipe gallery.
9. The method for positioning the abnormal position of the pipe rack based on the pipe rack video and the two-dimensional code detection of claim 8 is characterized in that in the hyper-diversity network:
connecting the network shallow features with the deep features by using a tight connection mode;
and a residual error structure is arranged for accelerating network convergence.
10. The utility model provides a piping lane abnormal position positioning system based on piping lane video and two-dimensional code detect which characterized in that includes:
the pipe gallery abnormity classification result acquisition module is used for inputting the picture of the pipe gallery to be positioned and identified into a pre-trained pipe gallery abnormity classification neural network model to acquire a pipe gallery abnormity classification result;
the two-dimensional code position acquisition module is used for inputting the picture of the pipe gallery to be positioned and identified into a pre-trained two-dimensional code detection model when the abnormal classification result of the pipe gallery is abnormal, and acquiring the position of the two-dimensional code in the picture of the pipe gallery to be positioned and identified;
the identification positioning module is used for acquiring and identifying a two-dimensional code picture according to the position of the two-dimensional code to acquire two-dimensional code pre-recorded information; based on the two-dimensional code is recorded information in advance and is accomplished piping lane abnormal position location.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110572994.4A CN113469937A (en) | 2021-05-25 | 2021-05-25 | Pipe gallery abnormal position positioning method and system based on pipe gallery video and two-dimensional code detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110572994.4A CN113469937A (en) | 2021-05-25 | 2021-05-25 | Pipe gallery abnormal position positioning method and system based on pipe gallery video and two-dimensional code detection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113469937A true CN113469937A (en) | 2021-10-01 |
Family
ID=77871540
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110572994.4A Pending CN113469937A (en) | 2021-05-25 | 2021-05-25 | Pipe gallery abnormal position positioning method and system based on pipe gallery video and two-dimensional code detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113469937A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108616723A (en) * | 2018-04-20 | 2018-10-02 | 国网江苏省电力有限公司电力科学研究院 | A kind of video routing inspection system for GIL piping lanes |
| CN108761290A (en) * | 2018-06-28 | 2018-11-06 | 浙江国自机器人技术有限公司 | Robot and its piping lane electric compartment method for inspecting, system, equipment, storage medium |
| CN109815765A (en) * | 2019-01-21 | 2019-05-28 | 东南大学 | A kind of method and device for extracting the business license information containing two dimensional code |
| CN110147864A (en) * | 2018-11-14 | 2019-08-20 | 腾讯科技（深圳）有限公司 | The treating method and apparatus of coding pattern, storage medium, electronic device |
| CN112165600A (en) * | 2020-08-26 | 2021-01-01 | 苏宁云计算有限公司 | Drowning identification method and device, camera and computer system |
| CN112347805A (en) * | 2020-11-25 | 2021-02-09 | 广东开放大学（广东理工职业学院） | Multi-target two-dimensional code detection and identification method, system, device and storage medium |
| CN112766795A (en) * | 2021-01-29 | 2021-05-07 | 长兴云尚科技有限公司 | Pipe gallery intelligent information management method and system based on cloud processing |
-
2021
- 2021-05-25 CN CN202110572994.4A patent/CN113469937A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108616723A (en) * | 2018-04-20 | 2018-10-02 | 国网江苏省电力有限公司电力科学研究院 | A kind of video routing inspection system for GIL piping lanes |
| CN108761290A (en) * | 2018-06-28 | 2018-11-06 | 浙江国自机器人技术有限公司 | Robot and its piping lane electric compartment method for inspecting, system, equipment, storage medium |
| CN110147864A (en) * | 2018-11-14 | 2019-08-20 | 腾讯科技（深圳）有限公司 | The treating method and apparatus of coding pattern, storage medium, electronic device |
| CN109815765A (en) * | 2019-01-21 | 2019-05-28 | 东南大学 | A kind of method and device for extracting the business license information containing two dimensional code |
| CN112165600A (en) * | 2020-08-26 | 2021-01-01 | 苏宁云计算有限公司 | Drowning identification method and device, camera and computer system |
| CN112347805A (en) * | 2020-11-25 | 2021-02-09 | 广东开放大学（广东理工职业学院） | Multi-target two-dimensional code detection and identification method, system, device and storage medium |
| CN112766795A (en) * | 2021-01-29 | 2021-05-07 | 长兴云尚科技有限公司 | Pipe gallery intelligent information management method and system based on cloud processing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112115927B (en) | Intelligent machine room equipment identification method and system based on deep learning | |
| CN111899227A (en) | Automatic railway fastener defect acquisition and identification method based on unmanned aerial vehicle operation | |
| CN116363125B (en) | Deep learning-based battery module appearance defect detection method and system | |
| CN112349057A (en) | Deep learning-based indoor smoke and fire detection method | |
| CN117541534A (en) | Power transmission line inspection method based on unmanned plane and CNN-BiLSTM model | |
| CN115171045A (en) | YOLO-based power grid operation field violation identification method and terminal | |
| CN113469938B (en) | Pipe gallery video analysis method and system based on embedded front-end processing server | |
| CN112766795B (en) | Cloud processing-based pipe gallery intelligent information management method and system | |
| CN110738272A (en) | method for labeling visualized mechanical continuous alarm samples of power transmission line channel | |
| CN110991300B (en) | Automatic identification method for abnormal swelling state of dorking abdomen | |
| CN117114420A (en) | Image recognition-based industrial and trade safety accident risk management and control system and method | |
| CN113297408A (en) | Image matching and scene recognition system and method based on Sift algorithm | |
| CN113469937A (en) | Pipe gallery abnormal position positioning method and system based on pipe gallery video and two-dimensional code detection | |
| CN111913619A (en) | Fire-fighting data reading and interaction method based on BIM lightweight | |
| CN116055521A (en) | Inspection system and image recognition method for electric inspection robot | |
| CN109902647A (en) | The portable online Bird's Nest intelligent identification Method of one kind and system | |
| CN115273009A (en) | Road crack detection method and system based on deep learning | |
| CN112784914B (en) | Pipe gallery video intelligent attribute detection method and system based on cloud processing | |
| CN115376023A (en) | Cultivation area detection method based on deformation convolutional network, unmanned aerial vehicle and medium | |
| CN113628253A (en) | Method and system for accurately detecting individual health of animal and storage medium | |
| CN113850773A (en) | Detection method, device, equipment and computer readable storage medium | |
| CN114199992B (en) | Method and system for detecting corrosion of tank wall of oil storage tank | |
| CN112487966B (en) | Mobile vendor behavior recognition management system | |
| CN118115591B (en) | Power prediction model training method, photovoltaic power station operation and maintenance method and related devices | |
| CN117218436A (en) | Transformer appearance defect identification method based on oblique photographing, three-dimensional reconstruction and yolov5 algorithm |
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 |