CN111582187A - Automatic meter reading method based on mixed Gaussian background modeling - Google Patents
Automatic meter reading method based on mixed Gaussian background modeling Download PDFInfo
- Publication number
- CN111582187A CN111582187A CN202010393108.7A CN202010393108A CN111582187A CN 111582187 A CN111582187 A CN 111582187A CN 202010393108 A CN202010393108 A CN 202010393108A CN 111582187 A CN111582187 A CN 111582187A
- Authority
- CN
- China
- Prior art keywords
- reading
- pointer
- background modeling
- mixed gaussian
- instrument
- 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
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000001914 filtration Methods 0.000 claims abstract description 18
- 238000009826 distribution Methods 0.000 claims description 24
- 238000005259 measurement Methods 0.000 claims description 11
- 238000007781 pre-processing Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 4
- 230000007704 transition Effects 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000008569 process Effects 0.000 abstract description 9
- 238000005516 engineering process Methods 0.000 abstract description 2
- 230000005540 biological transmission Effects 0.000 description 7
- 238000007689 inspection Methods 0.000 description 5
- 238000011161 development Methods 0.000 description 4
- 230000018109 developmental process Effects 0.000 description 4
- 238000012423 maintenance Methods 0.000 description 3
- 230000000694 effects Effects 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 238000007667 floating Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 230000000737 periodic effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 230000004069 differentiation Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 239000000779 smoke Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
- -1 wind Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/277—Analysis of motion involving stochastic approaches, e.g. using Kalman filters
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/02—Recognising information on displays, dials, clocks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Astronomy & Astrophysics (AREA)
- Remote Sensing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an automatic meter reading method based on mixed Gaussian background modeling, which is used for recording data of various meters in the operation process of a power grid system and obtaining a scheduling basis for ensuring the safe and stable operation of the whole power grid system on the basis of analyzing the data. The method adopts the camera to acquire the video data of the instrument panel and preprocesses the data; performing background modeling on the instrument panel by using mixed Gaussian background modeling, and extracting a pointer area of the instrument by using a background difference technology; tracking and filtering the pointer target with wrong adjustment by using a kalman filter to the candidate pointer area; and calculating the current reading of the instrument according to the position of the pointer in the instrument panel. The invention can realize automatic reading of the pointer instrument of the power grid equipment, can simultaneously realize reading of a plurality of dials, has the operation efficiency meeting the real-time requirement, effectively reduces the dependence of panel board patrol on manual work, and greatly improves the efficiency and the accuracy of the panel board patrol process.
Description
Technical Field
The invention relates to an automatic meter reading method based on mixed Gaussian background modeling, and belongs to the technical field of power dispatching automation.
Background
In the period of the complete development of industrialization in China, various non-renewable resources are gradually exhausted, and various renewable resources are urgently needed to be developed; in addition, the dependence of each industry on electric power resources is gradually increasing with the progress of industrialization. Renewable electric resources (such as water resources, wind resources and solar resources) in China are mainly and intensively distributed in the middle and western regions, economically developed regions are intensively distributed in the eastern coastal regions, and the situation that the electric resources and the economic development are not uniformly distributed needs to be broken through, and the situation that the electric resources and the economic development are only distributed by using a large-capacity long-distance power transmission system can be realized. The adoption of high-voltage and ultrahigh-voltage power transmission systems is a main mode for realizing long-distance transmission of electric power energy. After the ultrahigh-voltage high-capacity power transmission line is greatly expanded, the maintenance of safe and stable operation of the ultrahigh-voltage power transmission line is a solid foundation for high-speed increase of economic construction in China, and the high-quality high-frequency inspection operation and maintenance is a necessary condition for ensuring stable operation of a large-scale power transmission system. The manual inspection is easily affected by factors such as terrain, personnel and weather, so that the inspection efficiency is low, the inspection effect is poor, the real-time performance is poor, and the manual inspection maintenance mode cannot meet the requirement of power utility development in a new era. A large number of pointer type instruments are installed on various devices in a power grid system, and the running states of power devices are detected through data of the instruments, so that safe and stable running of the power grid system is scheduled and controlled. Compared with a digital instrument, the pointer instrument has the characteristics of low price, simple structure, electromagnetic interference resistance, long service life and the like, and can adapt to various complex electromagnetic interference environments in a power grid system for a long time. At present, reading of equipment and meters in a power grid system mainly depends on a manual mode, the equipment in the power grid system is various, the quantity of the meters is large, the labor intensity of manual reading is high, and the precision and the efficiency of the manual reading can be gradually reduced along with the lapse of working time. In addition, in some places with electricity, high voltage and high temperature, the manual reading mode cannot be used by means of intelligent identification.
Disclosure of Invention
A meter automatic reading method based on mixed Gaussian background modeling comprises the following steps:
1. image preprocessing and denoising;
2. modeling a mixed Gaussian background;
3. kalman filtering tracking;
4. and (4) calculating the reading of the meter.
Further, the present invention comprises:
1) image pre-processing denoising
The harmonic mean filtering processing can be used for effectively reducing noise. Specifically, the gaussian template filtering employed is as follows:
we generate a two-dimensional convolution template by discretizing a two-dimensional gaussian function, typically using a size of 3 x 3. The gaussian template of 3 x 3 is shown below:
2) gaussian mixture background modeling
Normal distribution (Normal distribution) is a very important probability distribution in the fields of mathematics, physics, engineering, etc., and has a great influence on many aspects of statistics. If the random variable X obeys a mathematical expectation of mu and the standard deviation of sigma2Is recorded as:
X～N(μ,σ2)；
the probability density function is then:
the expected value μ of a normal distribution determines its position and the standard deviation σ determines the amplitude of the distribution. Modeling is carried out on each pixel point in the image by adopting 3-5 normal distributions, and the states of the pixel points are represented by the weighted summation of the Gaussian distributions, so that the condition that the image background has a plurality of states can be well described, for example, the scene has periodic light change. The mathematical description is shown below. W hereinjIs the mean of the distribution, η (x; θ)j) Is the covariance matrix of the jth distribution, and x is the color [ R, G, B ] representing the pixel]A three-dimensional vector.
3) Kalman filter tracking
The Kalman filter algorithm, which describes an efficient linear measurement update on the estimate and error variance, can be divided into two parts:
3.1 time update (prediction) equation, as shown in the following equation.
3.2 measure the update equation, as shown in the following equation.
Kk＝Pk -HT(HPk -HT+R)-1
Pk＝(I-KkH)Pk -。
The motion of the pointer in the instrument panel is uniform circular motion, and the pointer is variable speed motion from the speed perspective, and belongs to a nonlinear system; however, circular motion is a uniform motion in terms of angular velocity of circular motion, and belongs to a linear system, and we can describe the system by a formula, namely a state transition equation of the system, wherein ω is angular velocity, R is radius, and v is linear velocity.
ωk+1＝ωk
vk+1＝R*ωk+1。
4) Meter reading calculation
In the process of using the hybrid Gaussian background modeling difference to detect the pointer of the instrument panel of the electric equipment, the image background of the instrument panel is changed greatly due to the complex environment of the pointer of the instrument panel, and for example, light changes, moving objects, floating objects and the like can be detected as candidate pointer areas. The motion state of the pointer of the instrument is described by using circular motion and tracked by using a Kalman filter to obtain the track of each candidate pointer target, if the current point of one track coincides with the starting point on a time sequence, the current pointer target is considered to be found, and the current reading value of the instrument is calculated by using the following formula, wherein value is the current reading of the table, valuemax is the maximum measurement of the table, and ω 0, ω 1 and ω max respectively represent the angle of the starting position of the pointer, the angle of the current position and the angle corresponding to the maximum measurement range.
The invention has the advantages of
The invention provides an automatic meter reading method based on mixed Gaussian background modeling, which comprises the steps of acquiring real-time video data of a meter panel through a network camera, performing Gaussian smoothing on a video image to remove Gaussian noise in the image, reconstructing a background image of the meter panel by adopting the mixed Gaussian background modeling method, extracting a pointer area through a background difference algorithm, performing Kalman filtering tracking on the pointer area to obtain a motion track of a pointer on the meter panel, determining a real pointer object through the circumferential characteristic of the track, and finally calculating the real-time reading of the meter panel. The algorithm of the invention can realize automatic reading of the pointer instrument of the power grid equipment, can simultaneously realize reading of a plurality of dials, has the operation efficiency meeting the real-time requirement, effectively reduces the dependence of panel board patrol on manual work, and greatly improves the efficiency and accuracy of the panel board patrol process.
Drawings
FIG. 1 is a flow chart of the method for automatic reading of a mixed Gaussian background modeling meter of the present invention.
Detailed Description
Referring to fig. 1, one embodiment of the present invention is:
1) image pre-processing denoising
Since the image is disturbed by various noises during the generation, transmission and recording processes, noise reduction is a very important step before further performing the operations of feature extraction, object recognition, etc. In the process of aerial photography, due to the influence of imaging equipment in the environment, light, electricity and other factors, the acquired image contains impulse noise. Meanwhile, due to the influence of random factors such as smoke, fog, wind, cloud and the like, the imaging viewpoint changes complexly, so that Gaussian noise exists in the acquired image inevitably. Gaussian noise is a random noise, which refers to noise whose n-dimensional distribution obeys gaussian distribution. Because gaussian noise is very well processed in both the spatial and frequency domains, its model is commonly used in practice. The harmonic mean filter reduces sharp transition of image gradation by averaging pixels in the neighborhood of the filter. The gaussian noise is caused by the sharp transformation of the gray scale, so that the harmonic mean filtering processing can be used for effectively reducing the noise. By the image preprocessing technology, the contrast and the quality of the image are greatly improved, the image characteristics of the research object can be more obvious, and the fault defect identification is facilitated. The gaussian template filtering is as follows:
we generate a two-dimensional convolution template by discretizing a two-dimensional gaussian function, typically using a size of 3 x 3. The gaussian template of 3 x 3 is shown below:
2) gaussian mixture background modeling
The basis of video object analysis is to separate the object region of interest from the scene, and the common separation methods are motion segmentation and background differentiation. The typical motion segmentation is to assume a constant background image in time series, and to demarcate a difference part between the background image and the current image by a threshold value. This method is relatively fast and easy to implement in many applications, but problems arise when tracking multiple targets or when a target suddenly stops. Due to the inherent drawbacks of motion detection based on frame differences, one has turned the direction of research to background differences, although the temporal complexity of the algorithm is increased when updating the background.
Normal distribution (Normal distribution) is a very important probability distribution in the fields of mathematics, physics, engineering, etc., and has a great influence on many aspects of statistics. If the random variable X obeys a mathematical expectation of mu and the standard deviation of sigma2Is recorded as:
X～N(μ,σ2)；
the probability density function is then:
the expected value μ of a normal distribution determines its position and the standard deviation σ determines the amplitude of the distribution. Modeling is carried out on each pixel point in the image by adopting 3-5 normal distributions, and the states of the pixel points are represented by weighted summation of Gaussian distributions, so that the existence of a plurality of states of the image background can be well describedSuch as the presence of periodic light changes in the scene. The mathematical description is shown below. W hereinjIs the mean of the distribution, η (x; θ)j) Is the covariance matrix of the jth distribution, and x is the color [ R, G, B ] representing the pixel]A three-dimensional vector.
3) Kalman filter tracking
Kalman filtering (Kalman filtering) is an algorithm that uses a linear system state equation to optimally estimate the state of a system by inputting and outputting observation data through the system. The optimal estimation can also be seen as a filtering process, since the observed data includes the effects of noise and interference in the system. Data filtering is a data processing technique for removing noise and restoring true data, and Kalman filtering can estimate the state of a dynamic system from a series of data with measurement noise under the condition that measurement variance is known. Kalamn filtering is the optimal solution to the problem of linear filtering of discrete data, and this filter can be derived in solving the optimal estimate of a linear system that is susceptible to white gaussian noise interference. The following system of equations may be used to determine a linear system given a discrete time process. The Kalman filter algorithm, which describes an efficient linear measurement update on the estimate and error variance, can be divided into two parts:
3.1 time update (prediction) equation, as shown in the following equation.
3.2 measure the update equation, as shown in the following equation.
Kk＝Pk -HT(HPk -HT+R)-1
Pk＝(I-KkH)Pk -。
The motion of the pointer in the instrument panel is uniform circular motion, and the pointer is variable speed motion from the speed perspective, and belongs to a nonlinear system; however, circular motion is a uniform motion in terms of angular velocity of circular motion, and belongs to a linear system, and we can describe the system by a formula, namely a state transition equation of the system, wherein ω is angular velocity, R is radius, and v is linear velocity.
ωk+1＝ωk
vk+1＝R*ωk+1
4) Meter reading calculation
In the process of using the hybrid Gaussian background modeling difference to detect the pointer of the instrument panel of the electric equipment, the image background of the instrument panel is changed greatly due to the complex environment of the pointer of the instrument panel, and for example, light changes, moving objects, floating objects and the like can be detected as candidate pointer areas. The motion state of the pointer of the instrument is described by using circular motion and tracked by using a Kalman filter to obtain the track of each candidate pointer target, if the current point of one track coincides with the starting point on a time sequence, the current pointer target is considered to be found, and the current reading value of the instrument is calculated by using the following formula, wherein value is the current reading of the table, valuemax is the maximum measurement of the table, and ω 0, ω 1 and ω max respectively represent the angle of the starting position of the pointer, the angle of the current position and the angle corresponding to the maximum measurement range.
Claims (2)
1. A meter automatic reading method based on mixed Gaussian background modeling is characterized by comprising the following steps:
1) image preprocessing and denoising;
2) modeling a mixed Gaussian background;
3) kalman filtering tracking;
4) calculating the reading of the meter;
the 1) image preprocessing denoising comprises:
the harmonic mean filtering processing can be used for effectively reducing noise. Specifically, the gaussian template filtering employed is as follows:
by discretizing a two-dimensional gaussian function, a two-dimensional convolution template is generated, typically using a size of 3 x 3. The gaussian template of 3 x 3 is shown below:
the 2) mixed Gaussian background modeling comprises the following steps:
the normal distribution is noted as:
X～N(μ,σ2)；
the probability density function is noted as:
modeling each pixel point in the image by normal distribution:
in the formula wjIs the mean of the distribution, η (x; θ)j) Is the covariance matrix of the jth distribution, and x is the color [ R, G, B ] representing the pixel]A three-dimensional vector;
the 3 Kalman filtering tracking comprises:
3.1) time update (prediction) equation:
3.2) measurement update equation:
Kk＝Pk -HT(HPk -HT+R)-1
Pk＝(I-KkH)Pk -；
the state transition equation of the system:
ωk+1＝ωk
vk+1＝R*ωk+1
where ω is angular velocity, R is radius, and v is linear velocity.
2. The method for automatically reading a meter based on mixed Gaussian background modeling according to claim 1, characterized in that:
the 4) meter reading calculation comprises the following steps:
where value is the current reading of the table, value max is the maximum measurement of the table, and ω 0, ω 1, and ω max represent the angle of the start position of the pointer, the angle of the current position, and the angle corresponding to the maximum measurement range, respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010393108.7A CN111582187A (en) | 2020-05-11 | 2020-05-11 | Automatic meter reading method based on mixed Gaussian background modeling |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010393108.7A CN111582187A (en) | 2020-05-11 | 2020-05-11 | Automatic meter reading method based on mixed Gaussian background modeling |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111582187A true CN111582187A (en) | 2020-08-25 |
Family
ID=72124849
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010393108.7A Pending CN111582187A (en) | 2020-05-11 | 2020-05-11 | Automatic meter reading method based on mixed Gaussian background modeling |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111582187A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102025981A (en) * | 2010-12-23 | 2011-04-20 | 北京邮电大学 | Method for detecting foreground in monitoring video |
| CN103134571A (en) * | 2013-01-24 | 2013-06-05 | 宁波市计量测试研究院（宁波市衡器管理所） | Water meter automatic verification technology based on dynamic video image processing and pattern recognition algorithm |
| CN103164692A (en) * | 2012-12-03 | 2013-06-19 | 北京科技大学 | Special vehicle instrument automatic identification system and algorithm based on computer vision |
| US20140232649A1 (en) * | 2011-09-09 | 2014-08-21 | Movea | Method of controlling a cursor by measurements of the attitude of a pointer and pointer implementing said method |
| CN106529559A (en) * | 2016-12-30 | 2017-03-22 | 山东鲁能软件技术有限公司 | Pointer-type circular multi-dashboard real-time reading identification method |
| CN108960237A (en) * | 2018-06-29 | 2018-12-07 | 上海海事大学 | A kind of pointer type fuel level indicator Recognition of Reading method |
-
2020
- 2020-05-11 CN CN202010393108.7A patent/CN111582187A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102025981A (en) * | 2010-12-23 | 2011-04-20 | 北京邮电大学 | Method for detecting foreground in monitoring video |
| US20140232649A1 (en) * | 2011-09-09 | 2014-08-21 | Movea | Method of controlling a cursor by measurements of the attitude of a pointer and pointer implementing said method |
| CN103164692A (en) * | 2012-12-03 | 2013-06-19 | 北京科技大学 | Special vehicle instrument automatic identification system and algorithm based on computer vision |
| CN103134571A (en) * | 2013-01-24 | 2013-06-05 | 宁波市计量测试研究院（宁波市衡器管理所） | Water meter automatic verification technology based on dynamic video image processing and pattern recognition algorithm |
| CN106529559A (en) * | 2016-12-30 | 2017-03-22 | 山东鲁能软件技术有限公司 | Pointer-type circular multi-dashboard real-time reading identification method |
| CN108960237A (en) * | 2018-06-29 | 2018-12-07 | 上海海事大学 | A kind of pointer type fuel level indicator Recognition of Reading method |
Non-Patent Citations (6)
| Title |
|---|
| 伍健荣等: "一种基于Kalman滤波器的自适应背景建模改进算法", 《传感器与微系统》 * |
| 刘伟等: "自适应混合高斯建模的高效运动目标检测", 《中国图象图形学报》 * |
| 季建平等: "一种准确提取设备系统仪表指示的方法", 《仪表技术》 * |
| 李学聪等: "指针式仪表图像的六步预处理方法", 《电测与仪表》 * |
| 郭贵法等: "改进的同态滤波在指针式仪表图像预处理中的应用", 《广东工业大学学报》 * |
| 黄超群: "基于混合高斯模型和kalman滤波器的运动目标检测与跟踪", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6759475B2 (en) | Ship detection methods and systems based on multidimensional features of the scene | |
| Sohn et al. | Automatic powerline scene classification and reconstruction using airborne lidar data | |
| CN106356757B (en) | A kind of power circuit unmanned plane method for inspecting based on human-eye visual characteristic | |
| CN109325935B (en) | Power transmission line detection method based on unmanned aerial vehicle image | |
| CN111428748A (en) | Infrared image insulator recognition and detection method based on HOG characteristics and SVM | |
| Liu et al. | Surface defect detection of solar cells based on feature pyramid network and GA-faster-RCNN | |
| Wu et al. | Detection of foreign objects intrusion into transmission lines using diverse generation model | |
| CN114241364A (en) | Method for quickly calibrating foreign object target of overhead transmission line | |
| CN106780560A (en) | A kind of feature based merges the bionic machine fish visual tracking method of particle filter | |
| CN110766782A (en) | Large-scale construction scene real-time reconstruction method based on multi-unmanned aerial vehicle visual cooperation | |
| CN108732570A (en) | The nowcasting method of disastrous convection weather based on particle filter blending algorithm | |
| CN115421150B (en) | Method and system for calculating distance between ship and shore based on laser radar point cloud data | |
| CN112329764A (en) | Infrared dim target detection method based on TV-L1 model | |
| CN113920255B (en) | High-efficient mapping system based on point cloud data | |
| CN111582187A (en) | Automatic meter reading method based on mixed Gaussian background modeling | |
| Xinlong et al. | Research on infrared small target tracking method | |
| Zhao et al. | Real-time power line segmentation detection based on multi-attention with strong semantic feature extractor | |
| CN116452604A (en) | Complex substation scene segmentation method, device and storage medium | |
| CN104392465A (en) | Multi-core target tracking method based on D-S evidence theory information integration | |
| CN111795982A (en) | High-voltage cable joint internal defect detection method based on X-ray imaging technology | |
| Jia et al. | Research on Intelligent Monitoring Technology of Traffic Flow Based on Computer Vision | |
| Zhao et al. | Transmission Line Fault Identification Method based on Attention Mechanism and Deep Neural Network | |
| CN112561945B (en) | Dynamic background target tracking method | |
| Sun et al. | Image segmentation algorithm based on top-hat transformation | |
| Cao et al. | Research on Automatic Screening Algorithm of Electric Power Image Recognition Data Based on Artificial Intelligence |
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: 20200825 |
|
| RJ01 | Rejection of invention patent application after publication |