CN112184624A - Picture detection method and system based on deep learning - Google Patents
Picture detection method and system based on deep learning Download PDFInfo
- Publication number
- CN112184624A CN112184624A CN202010905964.6A CN202010905964A CN112184624A CN 112184624 A CN112184624 A CN 112184624A CN 202010905964 A CN202010905964 A CN 202010905964A CN 112184624 A CN112184624 A CN 112184624A
- Authority
- CN
- China
- Prior art keywords
- image
- deep learning
- abnormal
- detection
- network 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
- 238000001514 detection method Methods 0.000 title claims abstract description 94
- 238000013135 deep learning Methods 0.000 title claims abstract description 63
- 230000002159 abnormal effect Effects 0.000 claims abstract description 67
- 238000012545 processing Methods 0.000 claims abstract description 47
- 238000000034 method Methods 0.000 claims abstract description 39
- 238000004458 analytical method Methods 0.000 claims abstract description 26
- 230000007547 defect Effects 0.000 claims abstract description 26
- 238000012800 visualization Methods 0.000 claims abstract description 13
- 238000001914 filtration Methods 0.000 claims abstract description 10
- 238000005070 sampling Methods 0.000 claims abstract description 10
- 238000002372 labelling Methods 0.000 claims description 8
- 239000004568 cement Substances 0.000 claims description 7
- 239000003086 colorant Substances 0.000 claims description 7
- 238000004901 spalling Methods 0.000 claims description 6
- 238000012423 maintenance Methods 0.000 abstract description 9
- 230000006870 function Effects 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 230000000007 visual effect Effects 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 4
- 238000004891 communication Methods 0.000 description 3
- 238000010276 construction Methods 0.000 description 3
- 230000008878 coupling Effects 0.000 description 3
- 238000010168 coupling process Methods 0.000 description 3
- 238000005859 coupling reaction Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000010191 image analysis Methods 0.000 description 3
- 238000012216 screening Methods 0.000 description 3
- 206010063385 Intellectualisation Diseases 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000002354 daily effect Effects 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000000750 progressive effect Effects 0.000 description 2
- 239000000725 suspension Substances 0.000 description 2
- 108010001267 Protein Subunits Proteins 0.000 description 1
- 230000005856 abnormality Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000011897 real-time detection Methods 0.000 description 1
- 229910052709 silver Inorganic materials 0.000 description 1
- 239000004332 silver Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
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
-
- 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
- G06N3/045—Combinations of networks
-
- 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
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- 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
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Quality & Reliability (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the application discloses a picture detection method and a picture detection system based on deep learning, wherein the picture detection system is used for automatically reading image data of rail transit on a hard disk storage medium; the image detection system inputs the image data into a pre-trained deep learning network model, and uses a GPU (graphics processing unit) for analysis, so that each layer of the deep learning network model performs convolution processing and sampling filtering on the input image, the characteristics of the input image are extracted, and the detected abnormal region of the input image is determined as a target region; analyzing the position and the confidence coefficient of the target area by the detection frames with different sizes, and marking the defect type and the confidence coefficient on the original image; and the CPU stores the output result to a database and a local designated path, and sends the output result to the front end for result statistics and visualization processing. The method solves the analysis requirement of mass image data generated in the rail transit operation and maintenance link, and improves the working efficiency and the intelligent level.
Description
Technical Field
The embodiment of the application relates to the technical field of intelligent traffic, in particular to a picture detection method and system based on deep learning.
Background
Rail transit refers to a type of vehicle or transportation system in which operating vehicles need to travel on a particular rail. The most typical rail transit is a railway system consisting of conventional trains and standard railways. With the diversified development of train and railway technologies, rail transit is more and more types, and is not only distributed in long-distance land transportation, but also widely applied to medium-short distance urban public transportation. The common rail transit includes traditional railways (national railways, intercity railways and urban railways), subways, light rails and trams, and the novel rail transit includes a magnetic suspension rail system, a monorail system (straddle type rail system and suspension type rail system), a passenger automatic rapid transit system and the like.
With the increasing mileage of rail transit in China, the detection and maintenance requirements for corresponding infrastructure are continuously expanded. The rail transit system adopts a large number of fixed or movable video and image acquisition devices to acquire the daily state of a rail and peripheral facilities thereof. This type of video, image capture device produces a large amount of data each day. How to efficiently acquire images and videos reflecting appearance defects and abnormalities of equipment from mass data has great significance for safe and stable operation of a rail transit system. In fact, in a rail transit system, the utilization of the massive image data is very limited, and most of the massive images are idle after being collected and cannot be analyzed in time; limited key parts such as a track surface, a rail and the like, and the image acquired by the equipment is basically screened and searched for defects by manual screening. The input of the automatic image shooting and recording equipment cannot be effectively and fully utilized, a large amount of manpower is required to be allocated for the automatic image shooting and recording equipment to manually analyze and view images, and the strange phenomenon that the higher the automation degree is, the manpower does not decrease and rise reversely occurs. At present, the rail transit operation and maintenance image analysis and identification is still a manual visual screening due to the lack of effective automatic defect identification tools and methods, and the defects of the method are needless to say: low speed, low accuracy, easy fatigue of manpower and high personnel cost.
How to efficiently process massive image information generated by daily operation and maintenance of the rail transit, and really realize end-to-end automation and intellectualization of the operation and maintenance inspection from acquisition to analysis is an important subject for improving the intellectualization level of railways and rail transit systems in China.
Disclosure of Invention
Therefore, the embodiment of the application provides a picture detection method and system based on deep learning, which are specially customized for batch analysis of mass video image data of rail transit. The system automatically reads image data on a hard disk storage medium, uses a deep learning AI algorithm to carry out intelligent analysis, automatically searches predefined characteristic targets such as defects, abnormity and the like of positioning related equipment or areas from mass image data, automatically stores detection images, simultaneously releases image inspectors from heavy visual detection work through Web data visual display, and greatly improves the detection efficiency and accuracy.
In order to achieve the above object, the embodiments of the present application provide the following technical solutions:
according to a first aspect of embodiments of the present application, there is provided a picture detection method based on deep learning, the method including:
the picture detection system automatically reads the image data of the rail transit on the hard disk storage medium;
the image detection system inputs the image data into a pre-trained deep learning network model, and uses a GPU (graphics processing unit) for analysis, so that each layer of the deep learning network model performs convolution processing and sampling filtering on the input image, the characteristics of the input image are extracted, and the detected abnormal region of the input image is determined as a target region;
analyzing the position and the confidence coefficient of the target area by the detection frames with different sizes, marking the defect type and the confidence coefficient on the original image, and inputting the defect type and the confidence coefficient into a CPU (Central processing Unit) as an output result; the detection frame is a labeling frame which is displayed on the inferred image and is identified as an abnormal area after the deep learning network model infers a frame of image;
and the CPU stores the output result to a database and a local designated path, and sends the output result to the front end for result statistics and visualization processing so as to check the detection result and the statistical results of all image detections in real time on a front end interface.
Optionally, the image detection system inputs the image data to a pre-trained deep learning network model, and performs analysis using a GPU, including: and inputting the image data into a pre-trained deep learning network model, and analyzing the image data by utilizing a plurality of GPUs (graphics processing units) in parallel.
Optionally, the method further comprises: the abnormal position is indicated by the marking frames with different colors on the abnormal area, and the picture detection system automatically stores an original picture directory of the abnormal image and an abnormal image marking picture directory under the same directory so as to facilitate the user to search and compare.
Optionally, the abnormal region of the input image is determined according to one or more of the conditions of capital construction wall, cracks on the cement surface of the road surface, spalling, abnormal states of equipment parts along the traffic line and user predefined situations of abnormal positions.
According to a second aspect of the embodiments of the present application, there is provided a deep learning based picture detection system, the system including:
the data reading module is used for automatically reading the image data of the rail transit on the hard disk storage medium;
the image intelligent detection module is used for inputting the image data into a pre-trained deep learning network model, analyzing by using a GPU (graphics processing unit), so that each layer of the deep learning network model performs convolution processing and sampling filtering on the input image, extracting the characteristics of the input image, and determining the detected abnormal region of the input image as a target region;
the output module is used for analyzing the position and the confidence coefficient of the target area by the detection frames with different sizes, marking the defect type and the confidence coefficient on the original image and inputting the defect type and the confidence coefficient into the CPU as an output result; the detection frame is a labeling frame which is displayed on the inferred image and is identified as an abnormal area after the deep learning network model infers a frame of image;
and the data sending module is used for storing the output result to a database and a local designated path by the CPU and sending the output result to the front end for result statistics and visualization processing so as to enable the detection result and the statistical results of all image detection to be checked in real time on a front end interface.
Optionally, the image intelligent detection module is specifically configured to:
and inputting the image data into a pre-trained deep learning network model, and analyzing the image data by utilizing a plurality of GPUs (graphics processing units) in parallel.
Optionally, the system further comprises: the abnormal position is indicated by the marking frames with different colors on the abnormal area, and the picture detection system automatically stores an original picture directory of the abnormal image and an abnormal image marking picture directory under the same directory so as to facilitate the user to search and compare.
Optionally, the abnormal region of the input image is determined according to one or more of the conditions of capital construction wall, cracks on the cement surface of the road surface, spalling, abnormal states of equipment parts along the traffic line and user predefined situations of abnormal positions.
According to a third aspect of embodiments herein, there is provided an apparatus comprising: the device comprises a data acquisition device, a processor and a memory; the data acquisition device is used for acquiring data; the memory is to store one or more program instructions; the processor is configured to execute one or more program instructions to perform the method of any of the first aspect.
According to a fourth aspect of embodiments herein, there is provided a computer-readable storage medium having one or more program instructions embodied therein for performing the method of any of the first aspects.
In summary, the embodiment of the present application provides a method and a system for detecting a picture based on deep learning, which automatically read image data of rail transit on a hard disk storage medium through a picture detection system; the image detection system inputs the image data into a pre-trained deep learning network model, and uses a GPU (graphics processing unit) for analysis, so that each layer of the deep learning network model performs convolution processing and sampling filtering on the input image, the characteristics of the input image are extracted, and the detected abnormal region of the input image is determined as a target region; analyzing the position and the confidence coefficient of the target area by the detection frames with different sizes, marking the defect type and the confidence coefficient on the original image, and inputting the defect type and the confidence coefficient into a CPU (Central processing Unit) as an output result; the detection frame is a labeling frame which is displayed on the inferred image and is identified as an abnormal area after the deep learning network model infers a frame of image; and the CPU stores the output result to a database and a local designated path, and sends the output result to the front end for result statistics and visualization processing so as to check the detection result and the statistical results of all image detections in real time on a front end interface. By means of technologies such as artificial intelligence deep learning algorithm and web visualization, the method aims to meet the analysis requirement of mass image data generated in the operation and maintenance link of the rail transit system and improve the working efficiency and the intelligent level.
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 will be briefly described below. It should be apparent that the drawings in the following description are merely exemplary, and that other embodiments can be derived from the drawings provided by those of ordinary skill in the art without inventive effort.
The structures, ratios, sizes, and the like shown in the present specification are only used for matching with the contents disclosed in the specification, so that those skilled in the art can understand and read the present invention, and do not limit the conditions for implementing the present invention, so that the present invention has no technical significance, and any structural modifications, changes in the ratio relationship, or adjustments of the sizes, without affecting the functions and purposes of the present invention, should still fall within the scope of the present invention.
Fig. 1 is a schematic flowchart of a deep learning-based picture detection method according to an embodiment of the present disclosure;
fig. 2 is a schematic diagram of an embodiment of a deep learning based picture detection method according to an embodiment of the present application;
fig. 3 is a block diagram of a deep learning based picture detection system according to an embodiment of the present application.
Detailed Description
The present invention is described in terms of particular embodiments, other advantages and features of the invention will become apparent to those skilled in the art from the following disclosure, and it is to be understood that the described embodiments are merely exemplary of the invention and that it is not intended to limit the invention to the particular embodiments disclosed. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment of the application provides a picture detection method flow schematic diagram based on deep learning, and aims at analyzing mass video image data of rail transit in batch. The system automatically reads image data on a hard disk storage medium, uses a special deep learning AI algorithm to carry out intelligent analysis, automatically searches predefined characteristic targets such as defects, abnormity and the like of positioning related equipment or areas from mass image data, automatically stores detection images, simultaneously releases image inspectors from heavy visual detection work through Web data visual display, and greatly improves the detection efficiency and accuracy.
Fig. 1 shows a schematic flowchart of a deep learning-based picture detection method provided in an embodiment of the present application, including the following steps:
step 101: the picture detection system automatically reads image data of rail traffic on the hard disk storage medium.
Step 102: the image detection system inputs the image data into a pre-trained deep learning network model, and uses a GPU for analysis, so that each layer of the deep learning network model performs convolution processing and sampling filtration on the input image, the characteristics of the input image are extracted, and the detected abnormal region of the input image is determined as a target region.
Step 103: analyzing the position and the confidence coefficient of the target area by the detection frames with different sizes, marking the defect type and the confidence coefficient on the original image, and inputting the defect type and the confidence coefficient into a CPU (Central processing Unit) as an output result; the detection frame is a labeling frame which is displayed on the inferred image and is identified as an abnormal area after the deep learning network model infers a frame of image.
Step 104: and the CPU stores the output result to a database and a local designated path, and sends the output result to the front end for result statistics and visualization processing so as to check the detection result and the statistical results of all image detections in real time on a front end interface.
In a possible implementation, the picture detection system inputs the image data into a pre-trained deep learning network model, and performs analysis using a GPU, including: and inputting the image data into a pre-trained deep learning network model, and analyzing the image data by utilizing a plurality of GPUs (graphics processing units) in parallel.
In one possible embodiment, the method further comprises: the abnormal position is indicated by the marking frames with different colors on the abnormal area, and the picture detection system automatically stores an original picture directory of the abnormal image and an abnormal image marking picture directory under the same directory so as to facilitate the user to search and compare.
In one possible embodiment, the abnormal region of the input image is determined according to one or more of the user predefined conditions of a capital wall, a crack on a cement surface of a road surface, spalling, an abnormal state of a component of equipment along a traffic line, and an abnormal position.
The embodiment of the application is also based on B/S architecture design, the system interface is simple and efficient, a client can use the system by opening a browser of the system, complex program installation and environment configuration are not needed, and the system simultaneously supports Linux and Windows system environments.
The system of the embodiment of the application also adopts a containerization design, and according to different configuration conditions of the client computer, the system can be arranged on a centralized high-performance multi-GPU server or used as an AI (analog to digital) computing plug-in of a common office PC (personal computer), and is distributed on embedded edge AI computing equipment for multiple people to use simultaneously.
According to the embodiment of the application, the automatic, batch and intelligent analysis of the big data of the rail transit operation and maintenance image is realized by using an artificial intelligent deep learning image analysis technology. A special deep learning network is designed and trained aiming at a rail transit operation and maintenance detection scene, the method has the characteristics of high speed, high accuracy and the like, various predefined defects are positioned efficiently, and the accuracy and effectiveness of detection are ensured.
In order to make the image detection method based on deep learning provided by the embodiment of the present application clearer, further explanation is performed with reference to the intelligent analysis system based on deep learning shown in fig. 2.
In a first aspect, a front-end web interface for human-computer interaction. The intelligent analysis system based on deep learning provided by the embodiment of the application is designed based on a B/S architecture, so that an operator can input a corresponding image path through a Web visual interface, and the processes of step-by-step retrieval and automatic detection of subdirectories under the path can be realized. The detection result and the statistical result can be displayed on the interface in real time, so that the detection result and the statistical result are convenient to view and evaluate. The abnormal position is indicated by the marking frame with different colors on the detected abnormal image, and simultaneously, the system can automatically store two catalogues under the same catalogues: the abnormal image original image and the abnormal image annotation image are convenient for a user to search and compare.
In order to meet the requirement of high-efficiency analysis of mass data, the system allocates image data in a certain range to be processed by different processes, each process allocates the data to multiple threads at the same time, batch processing can be performed on the image data, the multithreading capability of a CPU (Central processing Unit) is fully used, each GPU can be fully used in the environment with multiple GPUs, and the overall analysis speed is increased.
The artificial intelligent deep learning image analysis can intelligently identify abnormal areas in the image, such as user predefined conditions of the capital construction wall, cracks and peeling of the surface of pavement cement, abnormal states of equipment parts along the traffic line, abnormal positions and the like. The system inputs image data into a pre-trained deep learning network, uses a GPU for analysis, performs convolution processing and sampling filtering on input images through each layer of a neural network, extracts the characteristics of the input images, detects the input images on a plurality of sizes, inputs the input images into a CPU as output results, analyzes the position and confidence coefficient of a target by detection frames with different corresponding sizes, and then marks the target on an original image.
According to the target to be detected, the module can be flexibly and transversely expanded, namely, multiple modules can analyze multiple different targets in parallel. Meanwhile, according to the configuration condition of client hardware, the module can be arranged on a centralized high-performance multi-GPU server, can also be taken as an AI computing plug-in of a common office PC, and is distributed on embedded edge AI computing equipment.
And in the third aspect, the data storage and file operation module is used for carrying out result statistics and visualization processing on the image data after the image file data is subjected to batch processing, detection results are stored in the database and a local specified path, and meanwhile, the detection results and the statistical results of all image detection can be checked in real time on a front-end interface.
In specific implementation, the inspection image big data batch analysis tool applied to the B/S framework can be used for processing mass image data acquired by a rail transit system. The image data collected on site does not need to be preprocessed or operated by professionals, the external hard disk is simply connected, and the system can directly search and read the image directory and find out the suspected defect map from mass data. Thereby relieving existing human testers from the heavy picking labor. In an interactive interface of the system, one part is an image path selection module, the other part is a real-time statistical result visualization module, and the other part is a real-time detection result viewing module. The measured processing speed can reach 15 ten thousand high-resolution images analyzed and processed per hour.
Therefore, the following benefits can be obtained by the embodiments of the present application:
first, automatic batch. The system does not need manual intervention in the work, only needs to input a path to be analyzed after a user starts the system, automatically reads and searches step by step, completes a series of work such as analysis, marking, storage, display and the like, and can continuously work unattended for 7-24 hours.
Second, high speed. The traditional manual screening efficiency is low, and 5-6 workers can only complete the analysis of 10 ten thousand pictures at most after working for 8 hours every day. After the method is used, the single machine processing speed can reach 15 ten thousand high-resolution pictures analyzed and processed per hour. The specific operation efficiency may be different due to different configuration specifications of hardware or different input images: 10000 images at 2k resolution were processed and run on a platform of Intel-E5 Silver +2 x 2080Ti, taking about 6 minutes. On the platform of Intel-i7-9800x +1080Ti, it took about 8 minutes.
Thirdly, it is accurate. The system analyzes and judges the abnormal position predefined by the user with the accuracy rate of more than 90 percent, has no missing detection defect and the over-killing rate of 0.03 percent.
In summary, the embodiment of the present application provides a picture detection method based on deep learning, which automatically reads image data of rail transit on a hard disk storage medium through a picture detection system; the image detection system inputs the image data into a pre-trained deep learning network model, and uses a GPU (graphics processing unit) for analysis, so that each layer of the deep learning network model performs convolution processing and sampling filtering on the input image, the characteristics of the input image are extracted, and the detected abnormal region of the input image is determined as a target region; analyzing the position and the confidence coefficient of the target area by the detection frames with different sizes, marking the defect type and the confidence coefficient on the original image, and inputting the defect type and the confidence coefficient into a CPU (Central processing Unit) as an output result; the detection frame is a labeling frame which is displayed on the inferred image and is identified as an abnormal area after the deep learning network model infers a frame of image; and the CPU stores the output result to a database and a local designated path, and sends the output result to the front end for result statistics and visualization processing so as to check the detection result and the statistical results of all image detections in real time on a front end interface. By means of technologies such as artificial intelligence deep learning algorithm and web visualization, the method aims to meet the analysis requirement of mass image data generated in the operation and maintenance link of the rail transit system and improve the working efficiency and the intelligent level.
Based on the same technical concept, an embodiment of the present application further provides a deep learning-based picture detection system, as shown in fig. 3, the system includes:
and the data reading module 301 is used for automatically reading the image data of the rail transit on the hard disk storage medium.
The image intelligent detection module 302 is configured to input the image data to a pre-trained deep learning network model, perform analysis using a GPU, so that each layer of the deep learning network model performs convolution processing and sampling filtering on an input image, extract features of the input image, and determine a detected abnormal region of the input image as a target region.
The output module 303 is configured to analyze the position and the confidence of the target region by using the detection frames with different sizes, label the defect type and the confidence on the original image, and input the defect type and the confidence as an output result to the CPU; the detection frame is a labeling frame which is displayed on the inferred image and is identified as an abnormal area after the deep learning network model infers a frame of image.
And the data sending module 304 is used for storing the output result to a database and a local designated path by the CPU, and sending the output result to the front end for result statistics and visualization processing so as to enable the detection result and the statistical results of all image detections to be checked in real time on the front end interface.
In a possible implementation manner, the picture intelligent detection module 302 is specifically configured to: and inputting the image data into a pre-trained deep learning network model, and analyzing the image data by utilizing a plurality of GPUs (graphics processing units) in parallel.
In one possible embodiment, the system further comprises: the abnormal position is indicated by the marking frames with different colors on the abnormal area, and the picture detection system automatically stores an original picture directory of the abnormal image and an abnormal image marking picture directory under the same directory so as to facilitate the user to search and compare.
In one possible embodiment, the abnormal region of the input image is determined according to one or more of the user predefined conditions of a capital wall, a crack on a cement surface of a road surface, spalling, an abnormal state of a component of equipment along a traffic line, and an abnormal position.
Based on the same technical concept, an embodiment of the present application further provides an apparatus, including: the device comprises a data acquisition device, a processor and a memory; the data acquisition device is used for acquiring data; the memory is to store one or more program instructions; the processor is configured to execute one or more program instructions to perform the method.
Based on the same technical concept, the embodiment of the present application also provides a computer-readable storage medium, wherein the computer-readable storage medium contains one or more program instructions, and the one or more program instructions are used for executing the method.
In the present specification, each embodiment of the method is described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. Reference is made to the description of the method embodiments.
It is noted that while the operations of the methods of the present invention are depicted in the drawings in a particular order, this is not a requirement or suggestion that the operations must be performed in this particular order or that all of the illustrated operations must be performed to achieve desirable results. Additionally or alternatively, certain steps may be omitted, multiple steps combined into one step execution, and/or one step broken down into multiple step executions.
Although the present application provides method steps as in embodiments or flowcharts, additional or fewer steps may be included based on conventional or non-inventive approaches. The order of steps recited in the embodiments is merely one manner of performing the steps in a multitude of orders and does not represent the only order of execution. When an apparatus or client product in practice executes, it may execute sequentially or in parallel (e.g., in a parallel processor or multithreaded processing environment, or even in a distributed data processing environment) according to the embodiments or methods shown in the figures. The terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, the presence of additional identical or equivalent elements in a process, method, article, or apparatus that comprises the recited elements is not excluded.
The units, devices, modules, etc. set forth in the above embodiments may be implemented by a computer chip or an entity, or by a product with certain functions. For convenience of description, the above devices are described as being divided into various modules by functions, and are described separately. Of course, in implementing the present application, the functions of each module may be implemented in one or more software and/or hardware, or a module implementing the same function may be implemented by a combination of a plurality of sub-modules or sub-units, and the like. The above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
Those skilled in the art will also appreciate that, in addition to implementing the controller as pure computer readable program code, the same functionality can be implemented by logically programming method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Such a controller may therefore be considered as a hardware component, and the means included therein for performing the various functions may also be considered as a structure within the hardware component. Or even means for performing the functions may be regarded as being both a software module for performing the method and a structure within a hardware component.
The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, classes, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
From the above description of the embodiments, it is clear to those skilled in the art that the present application can be implemented by software plus necessary general hardware platform. Based on such understanding, the technical solutions of the present application may be embodied in the form of a software product, which may be stored in a storage medium, such as a ROM/RAM, a magnetic disk, an optical disk, or the like, and includes several instructions for enabling a computer device (which may be a personal computer, a mobile terminal, a server, or a network device) to execute the method according to the embodiments or some parts of the embodiments of the present application.
The embodiments in the present specification are described in a progressive manner, and the same or similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. The application is operational with numerous general purpose or special purpose computing system environments or configurations. For example: personal computers, server computers, hand-held or portable devices, tablet-type devices, multiprocessor systems, microprocessor-based systems, set top boxes, programmable electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments that include any of the above systems or devices, and the like.
The above-mentioned embodiments are further described in detail for the purpose of illustrating the invention, and it should be understood that the above-mentioned embodiments are only illustrative of the present invention and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (10)
1. A picture detection method based on deep learning is characterized by comprising the following steps:
the picture detection system automatically reads the image data of the rail transit on the hard disk storage medium;
the image detection system inputs the image data into a pre-trained deep learning network model, and uses a GPU (graphics processing unit) for analysis, so that each layer of the deep learning network model performs convolution processing and sampling filtering on the input image, the characteristics of the input image are extracted, and the detected abnormal region of the input image is determined as a target region;
analyzing the position and the confidence coefficient of the target area by the detection frames with different sizes, marking the defect type and the confidence coefficient on the original image, and inputting the defect type and the confidence coefficient into a CPU (Central processing Unit) as an output result; the detection frame is a labeling frame which is displayed on the inferred image and is identified as an abnormal area after the deep learning network model infers a frame of image;
and the CPU stores the output result to a database and a local designated path, and sends the output result to the front end for result statistics and visualization processing so as to check the detection result and the statistical results of all image detections in real time on a front end interface.
2. The method of claim 1, wherein the photo detection system inputs the image data into a pre-trained deep learning network model, and the analysis is performed using a GPU, comprising:
and inputting the image data into a pre-trained deep learning network model, and analyzing the image data by utilizing a plurality of GPUs (graphics processing units) in parallel.
3. The method of claim 1, wherein the method further comprises: the abnormal position is indicated by the marking frames with different colors on the abnormal area, and the picture detection system automatically stores an original picture directory of the abnormal image and an abnormal image marking picture directory under the same directory so as to facilitate the user to search and compare.
4. The method of claim 1, wherein the abnormal region of the input image is determined based on one or more of a capital wall, a surface cement crack, spalling, an abnormal condition of a piece of equipment along a traffic line, a user-predefined condition of an abnormal location.
5. A system for detecting pictures based on deep learning, the system comprising:
the data reading module is used for automatically reading the image data of the rail transit on the hard disk storage medium;
the image intelligent detection module is used for inputting the image data into a pre-trained deep learning network model, analyzing by using a GPU (graphics processing unit), so that each layer of the deep learning network model performs convolution processing and sampling filtering on the input image, extracting the characteristics of the input image, and determining the detected abnormal region of the input image as a target region;
the output module is used for analyzing the position and the confidence coefficient of the target area by the detection frames with different sizes, marking the defect type and the confidence coefficient on the original image and inputting the defect type and the confidence coefficient into the CPU as an output result; the detection frame is a labeling frame which is displayed on the inferred image and is identified as an abnormal area after the deep learning network model infers a frame of image;
and the data sending module is used for storing the output result to a database and a local designated path by the CPU and sending the output result to the front end for result statistics and visualization processing so as to enable the detection result and the statistical results of all image detection to be checked in real time on a front end interface.
6. The system of claim 5, wherein the picture intelligent detection module is specifically configured to:
and inputting the image data into a pre-trained deep learning network model, and analyzing the image data by utilizing a plurality of GPUs (graphics processing units) in parallel.
7. The system of claim 5, wherein the system further comprises: the abnormal position is indicated by the marking frames with different colors on the abnormal area, and the picture detection system automatically stores an original picture directory of the abnormal image and an abnormal image marking picture directory under the same directory so as to facilitate the user to search and compare.
8. The system of claim 5, wherein the abnormal region of the input image is determined based on one or more of a capital wall, a surface cement crack, spalling, an abnormal condition of a traffic track or equipment component along the line, a user predefined condition of an abnormal position.
9. An apparatus, characterized in that the apparatus comprises: the device comprises a data acquisition device, a processor and a memory;
the data acquisition device is used for acquiring data; the memory is to store one or more program instructions; the processor, configured to execute one or more program instructions to perform the method of any of claims 1-4.
10. A computer-readable storage medium having one or more program instructions embodied therein for performing the method of any of claims 1-4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010905964.6A CN112184624A (en) | 2020-09-01 | 2020-09-01 | Picture detection method and system based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010905964.6A CN112184624A (en) | 2020-09-01 | 2020-09-01 | Picture detection method and system based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112184624A true CN112184624A (en) | 2021-01-05 |
Family
ID=73924073
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010905964.6A Pending CN112184624A (en) | 2020-09-01 | 2020-09-01 | Picture detection method and system based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112184624A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115391084A (en) * | 2022-10-27 | 2022-11-25 | 北京蔚领时代科技有限公司 | Intelligent solution method and system for cloud game abnormity |
| WO2023108412A1 (en) * | 2021-12-14 | 2023-06-22 | 深圳先进技术研究院 | Rapid recognition and quantification method for target protein |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108304807A (en) * | 2018-02-02 | 2018-07-20 | 北京华纵科技有限公司 | A kind of track foreign matter detecting method and system based on FPGA platform and deep learning |
| CN110059631A (en) * | 2019-04-19 | 2019-07-26 | 中铁第一勘察设计院集团有限公司 | The contactless monitoring defect identification method of contact net |
| CN111241343A (en) * | 2020-01-07 | 2020-06-05 | 西安电子科技大学 | Road information monitoring and analyzing detection method and intelligent traffic control system |
-
2020
- 2020-09-01 CN CN202010905964.6A patent/CN112184624A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108304807A (en) * | 2018-02-02 | 2018-07-20 | 北京华纵科技有限公司 | A kind of track foreign matter detecting method and system based on FPGA platform and deep learning |
| CN110059631A (en) * | 2019-04-19 | 2019-07-26 | 中铁第一勘察设计院集团有限公司 | The contactless monitoring defect identification method of contact net |
| CN111241343A (en) * | 2020-01-07 | 2020-06-05 | 西安电子科技大学 | Road information monitoring and analyzing detection method and intelligent traffic control system |
Non-Patent Citations (1)
| Title |
|---|
| 连少远: ""交通图像中车辆感兴趣区域检测提取算法的研究与实现"", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅱ辑》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2023108412A1 (en) * | 2021-12-14 | 2023-06-22 | 深圳先进技术研究院 | Rapid recognition and quantification method for target protein |
| CN115391084A (en) * | 2022-10-27 | 2022-11-25 | 北京蔚领时代科技有限公司 | Intelligent solution method and system for cloud game abnormity |
| CN115391084B (en) * | 2022-10-27 | 2023-01-24 | 北京蔚领时代科技有限公司 | Intelligent solution method and system for cloud game abnormity |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108037133B (en) | Intelligent electric power equipment defect identification method and system based on unmanned aerial vehicle inspection image | |
| Maeda et al. | Road damage detection using deep neural networks with images captured through a smartphone | |
| CN110110613B (en) | Track traffic abnormal personnel detection method based on motion recognition | |
| EP2697738B1 (en) | Method and system of rail component detection using vision technology | |
| CN103778681B (en) | A kind of vehicle-mounted highway cruising inspection system and data acquisition and disposal route | |
| CN202793389U (en) | On-line routing inspection device for high-speed rail overhead line system | |
| CN105905132A (en) | Intelligent rail state patrolling device and method | |
| CN112184625A (en) | Pavement defect identification method and system based on video deep learning | |
| CN104778245A (en) | Similar trajectory mining method and device on basis of massive license plate identification data | |
| CN112184624A (en) | Picture detection method and system based on deep learning | |
| CN104504377A (en) | Bus passenger crowding degree identification system and method | |
| CN104077568A (en) | High-accuracy driver behavior recognition and monitoring method and system | |
| CN110021167A (en) | A kind of highway cloud inspection system and data processing method | |
| Doycheva et al. | GPU-enabled pavement distress image classification in real time | |
| CN115439643A (en) | Road disease size analysis and management method based on monocular measurement | |
| CN111767913A (en) | Motor train unit vehicle fault dynamic image detection method based on deep learning | |
| CN109885605B (en) | Urban physical sign index model data preprocessing method | |
| CN113469938B (en) | Pipe gallery video analysis method and system based on embedded front-end processing server | |
| Wang et al. | Automated shape-based pavement crack detection approach | |
| CN202815869U (en) | Vehicle microcomputer image and video data extraction apparatus | |
| Kuang et al. | Review on machine learning-based defect detection of shield tunnel lining | |
| CN113392770A (en) | Typical violation behavior detection method and system for transformer substation operating personnel | |
| CN102768730B (en) | Interactive wear particle image annotation method | |
| CN112365495A (en) | Method for identifying illegal building based on remote sensing | |
| CN111369062B (en) | Vehicle dynamics index extraction method and accident risk value prediction method |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210105 |
|
| RJ01 | Rejection of invention patent application after publication |