CN110837825A - Meter identification and reading system based on embedded type - Google Patents
Meter identification and reading system based on embedded type Download PDFInfo
- Publication number
- CN110837825A CN110837825A CN201910987187.1A CN201910987187A CN110837825A CN 110837825 A CN110837825 A CN 110837825A CN 201910987187 A CN201910987187 A CN 201910987187A CN 110837825 A CN110837825 A CN 110837825A
- Authority
- CN
- China
- Prior art keywords
- meter
- reading
- yolo
- identification
- image
- 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
- 230000006870 function Effects 0.000 claims abstract description 25
- 238000013135 deep learning Methods 0.000 claims abstract description 7
- 238000012549 training Methods 0.000 claims abstract description 4
- 238000000034 method Methods 0.000 claims description 14
- 238000007689 inspection Methods 0.000 claims description 13
- 230000003068 static effect Effects 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 6
- 238000004422 calculation algorithm Methods 0.000 description 9
- 238000012545 processing Methods 0.000 description 9
- 230000005540 biological transmission Effects 0.000 description 7
- 238000001514 detection method Methods 0.000 description 7
- 238000011897 real-time detection Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 4
- 241000282414 Homo sapiens Species 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 238000011160 research Methods 0.000 description 3
- HPTJABJPZMULFH-UHFFFAOYSA-N 12-[(Cyclohexylcarbamoyl)amino]dodecanoic acid Chemical compound OC(=O)CCCCCCCCCCCNC(=O)NC1CCCCC1 HPTJABJPZMULFH-UHFFFAOYSA-N 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 239000003595 mist Substances 0.000 description 2
- OHKOGUYZJXTSFX-KZFFXBSXSA-N ticarcillin Chemical compound C=1([C@@H](C(O)=O)C(=O)N[C@H]2[C@H]3SC([C@@H](N3C2=O)C(O)=O)(C)C)C=CSC=1 OHKOGUYZJXTSFX-KZFFXBSXSA-N 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 238000004883 computer application Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012552 review Methods 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Arrangements For Transmission Of Measured Signals (AREA)
Abstract
The invention relates to an embedded meter identification and reading system, which comprises the following steps: (1) collecting a video stream; (2) judging the fog concentration grade of the meter; (3) meter identification: the TX1 is provided with a GPU module and calls a YOLO deep learning framework through a GPU; the YOLO generates a weight file and a configuration file through the collected meter picture set in a training way, and modifies a detector.c function in the YOLO so that the function is not repeatedly loaded after the necessary weight file and configuration file are loaded once, but the picture is circularly detected; (4) meter reading and data backup.
Description
Technical Field
The invention relates to the field of embedded systems, in particular to a meter identification and reading system based on an embedded meter.
Background
Since the advent of the single chip microcomputer in 1971, various microprocessors are continuously updated with the continuous development of electronic technology, and an embedded system is originally based on the single chip microcomputer and has a history of nearly 50 years. An embedded system is a short name of an embedded special computer system, is different from an operating system in a general sense [1], is specially developed for solving a certain problem or realizing a certain function, and is often produced in batch so as to meet the requirement of manufacturing a certain device. With the continuous progress of modern technologies, especially the continuous popularization and development of various intelligent devices in recent years, the demand for various embedded systems is also increasing. The most prominent characteristic of the embedded system is that the embedded system has a small volume and is convenient to integrate into various small-sized devices or carry into various systems. In particular, the appearance of integrated circuits and the development of chip technology have led to the continuous reduction of the size of various microprocessors, and the application field of embedded systems has become more and more extensive.
The embedded system has a crucial position in the field of intelligent systems, particularly intelligent robots, and various embedded systems with small volume, high efficiency and strong specificity are important components of the intelligent robots. The intelligent robot is small in size and rich in functions, and can complete a lot of work which cannot be completed by human beings or replace the human beings to complete some dangerous work, so that the research of the intelligent robot becomes the popular research field in recent years. The inspection of the intelligent robot of the transformer substation is an important aspect of the application of the intelligent robot and is mainly responsible for replacing human beings to carry out the work of transformer substation abnormity detection, meter identification, meter reading and the like.
Various meters exist in a transformer substation, the number of the meters is large, and the traditional mode is that the meters are found manually, the types of the meters are marked, and reading records are carried out. Because only rely on the manual work to read the table, statistical data information can waste a large amount of manpower and materials and inefficiency, also inconvenient manual operation when the weather is abominable or the peripheral environment of meter is comparatively dangerous moreover leads to being difficult to satisfy the demand of transformer substation more and more. The intelligent robot inspection system can rapidly complete meter identification and reading through effective meter identification and reading algorithms, can adapt to various dangerous environments, saves a large amount of manpower and material resources, and greatly improves the working efficiency of the transformer substation. The meter recognition is generally based on a deep learning framework, and classification recognition is completed through training of a large number of samples, while the meter reading calculation method is mostly based on Hough detection [2-4], and is generally divided into three steps: pointer region extraction, pointer positioning and pointer reading [5], meter identification and reading all need to be performed with complex operations, and the requirements on equipment performance are high. Most of traditional meter identification and reading systems are realized on the basis of large-scale equipment such as computers, the identification and reading accuracy is high, but the size is large, and the traditional meter identification and reading systems cannot be generally used for realizing real-time detection of a mobile inspection system. Aiming at the problem, the meter identification and reading system based on the TX1 embedded system is provided, the embedded system is matched with a quad-core CPU and is also provided with a GPU special for image processing, a deep learning framework can be quickly operated to complete meter identification, image processing is efficiently carried out to complete meter reading, the size is not larger than a flat plate, the embedded system is conveniently embedded into an intelligent robot inspection system, and the acquisition of real-time video stream and the meter identification and reading work are completed along with the robot inspection process. Compared with the traditional meter identification and reading system, the system has the advantages that the size is reduced, and the efficiency and the precision of meter identification and reading are improved.
[1] Scribble just, yang riches people, xiu guang embedded operating system reviews [ J ]. computer application research, 2000,17(11).
[2]Hao Z L,Chen X H,Hu J Q,et al.OpenCV-Based Automatic DetectionSystem for AutomobileMeter[J].Applied Mechanics and Materials,2014,615:149-152.
[3]Zhang Z,Chen G,Li J,et al.The Research on Digit RecognitionAlgorithm for Automatic Meter Reading System[C]//Intelligent Control&Automation.IEEE,2010.
[4]Jianlong G,Liang G,Yaoyu L V,et al.Pointer meter reading methodbased on improved ORB and Hough algorithm[J].Computer Engineering andApplications,2018.
[5] A pointer type instrument reading identification system [ J ] in the inspection robot, 2017(7).
Disclosure of Invention
Aiming at overcoming the defects of the prior art and solving the problem that the traditional meter identification and reading system is too large in size and is not easy to apply to the inspection of the transformer substation robot, the invention aims to provide the meter identification and reading system based on TX1, and the meter identification and reading algorithm is integrated on a TX1 embedded system, and the four-core CPU and the high-efficiency GPU are used for identifying and reading the meter of the transformer substation, so that the requirement of reducing the size and carrying the meter to the intelligent inspection robot system of the transformer substation is met. Therefore, the technical scheme adopted by the invention is as follows:
an embedded meter-based meter identification and reading system, comprising the following steps:
(1) acquisition of video streams
Calling an rtsp protocol in the robot inspection process through a vehicle-mounted camera by the video stream, and acquiring a real-time video stream of the transformer substation;
(2) meter fog density grade discrimination
And (4) pre-judging the fog concentration grade of the meter, and performing next-step meter identification and reading after judging that the fog concentration can meet corresponding requirements.
(3) Meter identification
The TX1 is provided with a GPU module and calls a YOLO deep learning framework through a GPU; the YOLO generates a weight file and a configuration file through the collected meter picture set in a training way, and modifies a detector.c function in the YOLO so that the function is not repeatedly loaded after the necessary weight file and configuration file are loaded once, but the picture is circularly detected;
(4) meter reading and data backup
The method comprises the steps that a meter image and a corresponding meter area of which the meter reading needs to be static are judged through YOLO (YOLO) whether the coordinate of the meter is not obviously changed to determine whether a robot is static or not, then the image is collected, the corresponding meter area is determined according to the collected image, the detected meter type and the detected coordinate information, a reference line is calibrated after the meter area is defined, a meter pointer is extracted, the deflection angle of the meter pointer is calculated, the meter reading range is determined according to the meter type, and therefore the meter reading is carried out according to the deflection angle and the corresponding reading range; after the meter reading is finished, the collected meter image, the meter type, the meter coordinate and the meter reading information need to be transmitted back to the host computer end at regular time for data backup.
The invention provides a novel small-volume embedded system applied to real-time identification and reading of a transformer substation meter by integrating a meter identification and reading algorithm into a TX1 system, solves the problem that the traditional meter identification and reading system is too large in volume and cannot be applied to a real-time inspection system, and provides a new method for the intelligent inspection robot of the transformer substation to identify and read the meter.
Drawings
FIG. 1: and (5) a system design idea flow chart.
FIG. 2: the meter identifies the real-time display image.
FIG. 3: meter reading image.
In the figure: FIG. a is a chart image; fig. b is a pointer extraction image; graph c is the reading result
FIG. 4: and transmitting the image by data.
In the figure: FIG. a is a client; fig. b is a server.
Detailed Description
At present, most of traditional meter identification and reading systems are based on large-scale equipment such as computers, meter identification and reading are carried out through the powerful computing capability and the image processing capability of the computers, but the system cannot be combined with a vehicle-mounted system or an intelligent robot due to the huge size, so that only single-image or fixed-view-angle video stream detection can be realized, and the system cannot be applied to a transformer substation inspection robot for carrying out real-time meter identification and reading.
Aiming at the problem that the traditional meter identification and reading system is too large in size, an embedded system is provided. The system comprises four steps: firstly, a vehicle-mounted camera of the intelligent robot is utilized to call an rtsp protocol to acquire real-time video stream of the transformer substation, secondly, whether the fog concentration level of the meter meets the requirements of identification and reading is judged, then, a YOLO deep learning framework is called to identify the meter, and finally, the meter reading is carried out and a WIFI module is utilized to transmit data back to a host side for backup.
In order to describe the technical solution of the present invention more clearly, the following further description is made of the specific implementation process of the present invention. The invention is realized by the following steps:
(1) acquisition of video streams
Since the processing module is to process the image matrix, the collected video stream is converted into the image matrix before the next operation is performed. The rtsp protocol family has the advantage of being able to control to video frames, so the method of OpenCV + FFmpeg calling the rtsp video stream is finally used for acquiring the video stream. The video stream collected by the method consists of images of one frame and one frame, so that TX1 can directly call opencv library functions to process each frame, the method is simple and convenient, and the efficiency of the system is greatly improved. TX1 has a quad-core CPU, 4G run memory and 8G swap memory, and therefore can open up multiple threads for various processing. If the video stream is collected by opening up a single sub-thread, then the video stream is displayed by opening up a single thread. This not only makes reasonable use of the various outstanding properties of TX1, but also greatly speeds up operation. The image acquisition module is mainly responsible for acquiring video streams and transmitting the acquired video streams to the processing module, which requires that the acquisition module has both acquisition and transmission functions. The vehicle-mounted camera is a common network camera and is provided with an Ethernet interface, and TX1 is also provided with the Ethernet interface, so that a 10/100/1000Mbps Ethernet is supported by directly connecting an image acquisition module through a network cable, the speed is high, and the received video stream can reach 60 fps.
(2) Meter fog density grade discrimination
Fogging can occur in the air or on the dial due to weather effects, which can have a significant effect on the meter readings. And different mist concentrations will have different effects on meter identification and reading. For example, when the fog concentration is small, meter identification and reading are hardly influenced and are almost the same as the fog-free condition, and the meter identification and reading can be directly carried out without other treatment; when the fog concentration is moderate, some meters can accurately identify and read, and the identification and reading results of some meters are not accurate any more, so that the image needs to be subjected to certain defogging treatment; when the fog concentration is high, meter identification and reading cannot be carried out at all, and even if a defogging algorithm is added, the effect is not achieved. Therefore, the system is required to firstly judge the fog concentration level of the collected image and then adopt different processing methods according to different fog concentration levels. Considering that even the same meter account screen ratio with different fog concentration can cause different influences on the reading, the influence of the distance is also considered when judging the fog concentration level. And when the concentration of the mist can meet the corresponding requirement, carrying out next-step representation identification and reading.
(3) Meter identification and display
The meter identification module adopts dark learning of darknet, namely YOLO, because darknet is a comparatively light-duty open source degree of depth learning frame based on C and CUDA completely, its key feature is that it is easy to install, there is not any dependence (OpenCV can all be unnecessary), the portability is very good, support CPU and GPU two kinds of calculation modes, and can reach higher recognition rate when realizing quick detection. CUDA multithreading concurrent processing of the GPU and four-core parallel operation of the CPU enable the TX1 to complete a large amount of calculation in a short time, and the image processing speed is greatly increased. And sufficient operation memory enables TX1 to be embedded into a YOLO deep learning framework, real-time detection of the meter is completed, and flexibility of the system is greatly improved. And TX1 is small in size and does not exceed the size of a flat plate overall, so that the system can be mounted on various vehicle-mounted systems, and the application range is wider.
YOLO trains itself to generate weight files (weights files) and configuration files (cfg files) through the collected tabulated picture sets. For video detection or real-time detection of camera video stream, YOLO gives a demo function, but the demo has higher requirement on hardware performance, so TX1 directly runs the demo and is very stuck, and therefore YOLO cannot be directly used for real-time detection, and finally YOLO source codes are modified for detection. The demo given by YOLO loads the weight file and the configuration file once when detecting each frame, which greatly reduces the operation efficiency. And modifying the detector.c function in the YOLO, so that the pictures are not repeatedly loaded after the necessary weight file and configuration file are loaded once, but are circularly detected, and finally, the detection of each picture only needs 400 and 500 ms. Even if the real-time detection and real-time display can be realized, if each frame is processed once, a great delay is caused, one frame is transmitted after the frame is processed, then the coordinates of the box are transmitted back to the master function, the yolo display function cannot be used, the frame is drawn on each frame according to the transmitted box coordinates before the master function is displayed, the recognized type character strings are transmitted back to the master function and are written into each frame together.
The representation is identified and then displayed in real time, and the TX1 is provided with an HDMI audio/video interface and a WiFi module. Therefore, the display module can be directly connected with a high-definition screen by adopting an HDMI wire, and can also be connected with TX1 for displaying by adopting mobile equipment such as a tablet, a mobile phone and the like through accessing the same local area network. This makes the whole system more flexible and easy to control.
Because Qt has the advantages of simplicity, easy operation, good portability and the like, the interface is finally designed by adopting Qt. And newly building a QtWidgets Application project, and finally selecting a self-written code to realize an interface because the UI editor of the Qt is too limited. The interface consists of a time part, a main body display part and a function key part. The time section displays the current time. The meter is displayed on the right side of the main body display part to identify the real-time video stream, and the coordinates of a meter identification frame and the meter category are displayed in real time in the lower label; the left side can display a certain captured frame of image, and the lower label displays information such as the coordinate, category and reading of a meter in the captured image; when the system is stationary for 1-2s, the system will capture a frame and save it and display it on the left side, and display some information of the meter in the image in label. The function key portion has four keys in total. The Camera key is used for starting a Camera and starting video stream acquisition and meter real-time detection; the Grab key is used for capturing a frame of picture and displaying the frame on the left side, and after the Grab captures the frame, the left display frame suspends capturing the picture when the system is static for 1-2 s; the ReImgshow key is used for resetting the left display frame after the Grab captures the frame, so that the frame can be captured and displayed when the system is static for 1-2 s; the Shutdown button is used to turn off TX 1.
(4) Meter reading and data backup
The meter reading needs a static meter image, so when the intelligent robot patrols the meter position, the intelligent robot is firstly static for 1-2s, the detected meter coordinate is judged by using a YOLO algorithm, and when the coordinate change is smaller than a certain threshold value, the system can be considered to be in a static state at the moment. The method comprises the steps of transmitting relevant information such as meter images acquired when a system is static, meter coordinates identified by YOLO, meter types and the like to a meter reading algorithm, and cutting the meter images according to the meter coordinates by the meter reading algorithm to ensure that the meter occupies a screen ratio enough to enable the reading to be more convenient and accurate. And then extracting a pointer region, in order to remove the influence of factors such as illumination and the like and enhance the robustness of the algorithm, denoising is required before the pointer region is extracted. And extracting the meter pointer based on a Hough algorithm after the pointer area is successfully extracted, calculating the deflection angle of the pointer by comparing the reference point, determining the reading range of the meter according to the type of the meter, and fitting the deflection angle of the pointer and the reference point with the dial scale according to the reading range of the meter to obtain the corresponding meter reading.
After obtaining the information such as the picture, box coordinate, meter type, and mark reading, the information needs to be transmitted back to the host computer for backup or further processing. The transmission module adopts a WIFI module of TX1 and adopts TCP/IP protocol for transmission. This is because the in-vehicle system is always moving and the host is stationary, so the system and the host cannot be connected by physical wiring and only wireless transmission can be used. The TX1 is provided with a WIFI module and can be connected with a WIFI network, so that wireless transmission can be carried out by adopting a TCP/IP protocol as long as the vehicle-mounted system and the host are in the same local area network. The specific implementation method is implemented by socket sockets. The socket () function contains three arguments. Wherein the AF parameter is an address description. Currently, only the AF _ INET format, i.e. the arpintenet address format, is supported. The Type parameter is used to specify a socket Type, and commonly used socket types are SOCK _ STREAM, SOCK _ DGRAM, and the like. Protocol parameters, as the name implies, are specific protocols. The protocol used by the socket. If the caller does not want to specify, 0 may be used. Common protocols include IPPROTO _ TCP, IPPROTO _ UDP, IPPROTO _ STCP, IPPROTO _ TIPC, etc., which correspond to TCP transport protocol, UDP transport protocol, STCP transport protocol, TIPC transport protocol, respectively. Here we type parameter select SOCK _ STREAM, protocol parameter default is to automatically select TCP protocol.
The transmission part is divided into a client and a server, and the system realizes that TX1 transmits to a host, so TX1 serves as the client and the host serves as the server. Firstly, socket () functions are used for creating sockets at a server and a client respectively, and bind the sockets by using a bind () function, then the server needs to call a list () function to enter a monitoring stage, whether the client sends a connection request is monitored, the client calls a connect () function to send the connection request, and at the moment, the server is connected to the connection request of the client and then calls an accept () function to accept connection. Thus, reliable connection is established, and transmission quality and stability are guaranteed. When a file is transferred, file operation functions such as fopen (), fread (), and fwrite () need to be called, and the file is opened, read, and written in a binary form.
Claims (1)
1. An embedded meter-based meter identification and reading system, comprising the following steps:
(1) acquisition of video streams
Calling an rtsp protocol in the robot inspection process through a vehicle-mounted camera by the video stream, and acquiring a real-time video stream of the transformer substation;
(2) meter fog density grade discrimination
And (4) pre-judging the fog concentration grade of the meter, and performing next-step meter identification and reading after judging that the fog concentration can meet corresponding requirements.
(3) Meter identification
The TX1 is provided with a GPU module and calls a YOLO deep learning framework through a GPU; the YOLO generates a weight file and a configuration file through the collected meter picture set in a training way, and modifies a detector.c function in the YOLO so that the function is not repeatedly loaded after the necessary weight file and configuration file are loaded once, but the picture is circularly detected;
(4) meter reading and data backup
The method comprises the steps that a meter image and a corresponding meter area of which the meter reading needs to be static are judged through YOLO (YOLO) whether the coordinate of the meter is not obviously changed to determine whether a robot is static or not, then the image is collected, the corresponding meter area is determined according to the collected image, the detected meter type and the detected coordinate information, a reference line is calibrated after the meter area is defined, a meter pointer is extracted, the deflection angle of the meter pointer is calculated, the meter reading range is determined according to the meter type, and therefore the meter reading is carried out according to the deflection angle and the corresponding reading range; after the meter reading is finished, the collected meter image, the meter type, the meter coordinate and the meter reading information need to be transmitted back to the host computer end at regular time for data backup.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910987187.1A CN110837825A (en) | 2019-10-17 | 2019-10-17 | Meter identification and reading system based on embedded type |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910987187.1A CN110837825A (en) | 2019-10-17 | 2019-10-17 | Meter identification and reading system based on embedded type |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110837825A true CN110837825A (en) | 2020-02-25 |
Family
ID=69575601
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910987187.1A Pending CN110837825A (en) | 2019-10-17 | 2019-10-17 | Meter identification and reading system based on embedded type |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110837825A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112308054A (en) * | 2020-12-29 | 2021-02-02 | 广东科凯达智能机器人有限公司 | Automatic reading method of multifunctional digital meter based on target detection algorithm |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005267365A (en) * | 2004-03-19 | 2005-09-29 | Ntt Power & Building Facilities Inc | Method for reading pointer value of pointer rotary meter, device for reading pointer value of pointer rotary meter and program for reading meter pointer value |
| CN102202163A (en) * | 2011-05-13 | 2011-09-28 | 成都西图科技有限公司 | Adaptive enhancement method and device for monitored video |
| CN104714068A (en) * | 2015-04-01 | 2015-06-17 | 无锡桑尼安科技有限公司 | Robot for outdoor electric energy meter to read meter |
| CN107139179A (en) * | 2017-05-26 | 2017-09-08 | 西安电子科技大学 | A kind of intellect service robot and method of work |
| CN108665432A (en) * | 2018-05-18 | 2018-10-16 | 百年金海科技有限公司 | A kind of single image to the fog method based on generation confrontation network |
| CN109543682A (en) * | 2018-11-23 | 2019-03-29 | 电子科技大学 | A kind of readings of pointer type meters method based on deep learning |
| CN109584186A (en) * | 2018-12-25 | 2019-04-05 | 西北工业大学 | A kind of unmanned aerial vehicle onboard image defogging method and device |
| CN109816610A (en) * | 2019-01-31 | 2019-05-28 | 上海工程技术大学 | A kind of image defogging system |
| CN109948469A (en) * | 2019-03-01 | 2019-06-28 | 吉林大学 | The automatic detection recognition method of crusing robot instrument based on deep learning |
| CN109993154A (en) * | 2019-04-19 | 2019-07-09 | 山东大学 | The lithium sulfur type instrument intelligent identification Method of substation's simple pointer formula |
| CN110321855A (en) * | 2019-07-07 | 2019-10-11 | 徐梓恒 | A kind of greasy weather detection prior-warning device |
-
2019
- 2019-10-17 CN CN201910987187.1A patent/CN110837825A/en active Pending
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005267365A (en) * | 2004-03-19 | 2005-09-29 | Ntt Power & Building Facilities Inc | Method for reading pointer value of pointer rotary meter, device for reading pointer value of pointer rotary meter and program for reading meter pointer value |
| CN102202163A (en) * | 2011-05-13 | 2011-09-28 | 成都西图科技有限公司 | Adaptive enhancement method and device for monitored video |
| CN104714068A (en) * | 2015-04-01 | 2015-06-17 | 无锡桑尼安科技有限公司 | Robot for outdoor electric energy meter to read meter |
| CN107139179A (en) * | 2017-05-26 | 2017-09-08 | 西安电子科技大学 | A kind of intellect service robot and method of work |
| CN108665432A (en) * | 2018-05-18 | 2018-10-16 | 百年金海科技有限公司 | A kind of single image to the fog method based on generation confrontation network |
| CN109543682A (en) * | 2018-11-23 | 2019-03-29 | 电子科技大学 | A kind of readings of pointer type meters method based on deep learning |
| CN109584186A (en) * | 2018-12-25 | 2019-04-05 | 西北工业大学 | A kind of unmanned aerial vehicle onboard image defogging method and device |
| CN109816610A (en) * | 2019-01-31 | 2019-05-28 | 上海工程技术大学 | A kind of image defogging system |
| CN109948469A (en) * | 2019-03-01 | 2019-06-28 | 吉林大学 | The automatic detection recognition method of crusing robot instrument based on deep learning |
| CN109993154A (en) * | 2019-04-19 | 2019-07-09 | 山东大学 | The lithium sulfur type instrument intelligent identification Method of substation's simple pointer formula |
| CN110321855A (en) * | 2019-07-07 | 2019-10-11 | 徐梓恒 | A kind of greasy weather detection prior-warning device |
Non-Patent Citations (1)
| Title |
|---|
| 张珺: "基于机器视觉的仪表读数智能识别算法研究及应用", pages 3 - 4 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112308054A (en) * | 2020-12-29 | 2021-02-02 | 广东科凯达智能机器人有限公司 | Automatic reading method of multifunctional digital meter based on target detection algorithm |
| CN112308054B (en) * | 2020-12-29 | 2021-07-20 | 广东科凯达智能机器人有限公司 | Automatic reading method of multifunctional digital meter based on target detection algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112633384B (en) | Object recognition method and device based on image recognition model and electronic equipment | |
| CN104320615B (en) | A kind of intelligent video safety-protection system and its signal processing method | |
| CN112200187A (en) | Target detection method, device, machine readable medium and equipment | |
| CN203629603U (en) | Intelligent electric energy meter detection device based on image identification technology | |
| CN108952673B (en) | Method and device for checking working condition of oil pumping well | |
| CN110969604B (en) | Intelligent security real-time windowing detection alarm system and method based on deep learning | |
| US20220323999A1 (en) | Method and apparatus for detecting wafer cleaning anomalies | |
| WO2017113874A1 (en) | Board card error detection method and system | |
| CN110837825A (en) | Meter identification and reading system based on embedded type | |
| CN111401206A (en) | Panorama sharing method, system, device and medium | |
| CN204479457U (en) | The device that colourimetry detects fast | |
| CN102023163B (en) | System and method for detecting connector based on digital signal processor (DSP) | |
| CN112199997A (en) | Terminal and tool processing method | |
| CN116884037A (en) | Image recognition method, device, equipment and storage medium | |
| CN104914985A (en) | Gesture control method and system and video flowing processing device | |
| CN114173158B (en) | Face recognition method, cloud device, client device, electronic device and medium | |
| CN115984694A (en) | Intelligent inspection method, device, system and equipment for urban house illegal construction | |
| CN113591752B (en) | Equipment oil leakage monitoring method based on convolutional neural network and related equipment | |
| US20240071023A1 (en) | Method and apparatus for detecting near-field object, and medium and electronic device | |
| CN105573605A (en) | Interactive game method of mobile phone and computer | |
| CN210052147U (en) | Mobile intelligent terminal screen controller | |
| CN107656816A (en) | Multi-terminal synchronous operation method and device | |
| CN101576788B (en) | Multi-point touch positioning method | |
| CN111626369A (en) | Face recognition algorithm effect evaluation method and device, machine readable medium and equipment | |
| CN105573688A (en) | Multi-screen interoperation method based on image capture |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200225 |