CN114821452A - Colored drawing train number identification method, system and medium - Google Patents
Colored drawing train number identification method, system and medium Download PDFInfo
- Publication number
- CN114821452A CN114821452A CN202210744590.3A CN202210744590A CN114821452A CN 114821452 A CN114821452 A CN 114821452A CN 202210744590 A CN202210744590 A CN 202210744590A CN 114821452 A CN114821452 A CN 114821452A
- Authority
- CN
- China
- Prior art keywords
- image
- train
- colored drawing
- character
- train number
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- 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
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/14—Image acquisition
- G06V30/148—Segmentation of character regions
- G06V30/153—Segmentation of character regions using recognition of characters or words
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V30/00—Character recognition; Recognising digital ink; Document-oriented image-based pattern recognition
- G06V30/10—Character recognition
- G06V30/16—Image preprocessing
- G06V30/164—Noise filtering
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Train Traffic Observation, Control, And Security (AREA)
Abstract
The invention relates to a colored drawing train number identification method, a colored drawing train number identification system and a colored drawing train number identification medium, and belongs to the technical field of image processing; the colored drawing train number identification method comprises the steps of S1, obtaining a train video data stream collected by monitoring, obtaining inter-frame information through the video data stream, and obtaining an image of a train movement change area according to the inter-frame information; s2, preprocessing the image, and removing noise interference through Gaussian filtering; s3, positioning the car number character region of the image without noise interference to obtain a segmentation image containing the car number character region; s4, performing HSL conversion processing on the segmentation image containing the car number character area; s5, performing character recognition on the image subjected to the HSL conversion processing through a lightweight CRNN network model; the invention realizes automatic recognition of the train number of the colored drawing train, improves the accuracy of train number recognition, reduces the time of train number recognition and has high automation degree.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a colored drawing train number identification method, a colored drawing train number identification system and a colored drawing train number identification medium.
Background
At present, urbanization is more and more deep, the popularity of trains is increased, the safety problem of the trains is enhanced, the trains can be overhauled when stopped, and the train numbers of the trains need to be acquired when the trains are overhauled.
With the development of the technology, the propaganda of local culture is promoted, and colored drawing patterns appear on the train body. Such as pelican, pteridium aquilinum, kapok, hummingbird, etc. The colored drawing patterns shield part of the car numbers on the car body, and the colors of some colored drawing patterns are similar to the colors of the car numbers, so that the difficulty of car number identification is greatly increased.
Disclosure of Invention
In order to solve the problems in the background art, the invention provides a method, a system and a medium for identifying a colored drawing train number.
In a first aspect, the invention provides a colored drawing train number identification method, which adopts the following technical scheme:
a colored drawing train number identification method comprises the following steps:
acquiring a train video data stream collected by monitoring, acquiring interframe information through the video data stream, and acquiring an image of a train motion change area according to the interframe information;
preprocessing an image to remove noise interference;
positioning the car number character region of the image without noise interference to obtain a segmented image containing the car number character region;
performing HSL transformation processing on the segmentation image;
and performing character recognition on the image subjected to the HSL conversion processing.
By adopting the technical scheme, the video data stream of the train collected by monitoring is obtained, the area of the movement change of the train is obtained according to the interframe information to obtain the image of the train, the image is preprocessed to remove noise interference, the definition and the quality of the image are improved, the image without the noise interference is positioned in the character area of the train number and is intercepted to obtain the segmentation image containing the character area of the train number, the segmentation image is subjected to HSL conversion to obtain the brightness of the image, the higher the brightness value is, the whiter the color is, the lower the brightness value is, the blacker the color is, the color different from the black train number in the segmentation image is removed according to the brightness value, the black character image of the train number is left for identification, the automation degree is high, the difficulty of the identification of the train number is reduced, the train number of the train can be effectively identified, the accuracy of the identification of the train number is improved, and the time of the identification of the train number is reduced, the efficiency of train maintenance is improved.
Preferably, the image is preprocessed to remove noise interference specifically as follows:
and removing noise interference by Gaussian filtering by adopting a filtering algorithm. By adopting the filtering algorithm, noise interference can be effectively removed, the definition and the quality of the image are improved, and subsequent work such as target positioning, identification and the like is facilitated.
Preferably, the car number character region positioning is performed on the image without the noise interference to obtain a segmented image containing the car number character region, and the steps specifically include:
acquiring an image without noise interference, and sending the image to a preset character area positioning model;
performing car number character area positioning on the image without noise interference by using the preset character area positioning model;
and segmenting the image of the positioned car number character area to obtain a segmented image containing the car number character area.
Preferably, the specific generation steps of the character region positioning model are as follows:
acquiring an initial model and a first data set, wherein the first data set comprises a conventional train picture and a colored drawing train picture;
carrying out region labeling on characters in the train number picture of the train to form a first training set and a first testing set;
and performing target detection network training on the initial model according to the first training set and the first testing set to generate a character area positioning model.
By segmenting the image of the positioned vehicle number character region, the influence of the non-vehicle number character region on character recognition is reduced, and the accuracy of vehicle number recognition is improved.
Preferably, the performing HSL transform processing on the segmented image specifically includes:
the method comprises the steps of performing HSL conversion on a divided image containing a car number character area according to the divided image, and processing the divided image according to L information in the HSL conversion.
By performing the HSL conversion on the image, the brightness L of the image, i.e., the brightness of the color, can be obtained, the higher the brightness value, the whiter the color, the lower the brightness value, and the blacker the color, and the color different from the black car number in the divided image can be removed according to the obtained brightness value, leaving a black car number character image.
Preferably, the character recognition of the image after the HSL conversion processing includes:
acquiring an image subjected to HSL conversion processing, and sending the image to a pre-generated lightweight CRNN network model;
and recognizing characters in the image subjected to the HSL conversion processing by using the pre-generated lightweight CRNN network model.
By adopting the lightweight CRNN network model, the model becomes smaller after the model quantization, and the identification time is shortened.
Preferably, the specific generation steps of the lightweight CRNN network model are as follows:
acquiring a second data set, wherein the second data set is a conventional train picture and a colored drawing train picture;
processing the second data set through a character area positioning model to obtain a segmentation image containing a car number character area, performing HSL (high speed Link) conversion on the segmentation image, and processing the segmentation image by using brightness L information to obtain an image to form a second training set and a second test set;
and carrying out network training according to the second training set and the second testing set to generate a lightweight CRNN network model.
In a second aspect, the invention provides a system for identifying a train number of a colored drawing train, which adopts the following technical scheme:
a system for colored drawing train number identification, comprising:
the video acquisition module is used for acquiring video data of train running, acquiring interframe information through the video data and acquiring an image of a train motion change area according to the interframe information;
the noise removing module is used for preprocessing the acquired images of the train motion change area and removing noise interference;
the character positioning module is used for positioning the car number character region of the image without the noise interference to obtain a segmented image containing the car number character region;
an HSL conversion module for performing HSL conversion processing on the segmentation image based on the region containing the car number characters;
and the character recognition module is used for carrying out character recognition on the image subjected to the HSL conversion processing.
By adopting the technical scheme, the video data of the train is acquired through the video acquisition module, the image of the train is acquired according to the inter-frame information, the noise interference is removed through the preprocessing of the image by the noise removal module, the definition and the quality of the image are improved, the image without the noise interference is positioned in the train number character region through the character positioning module and is intercepted, the segmentation image containing the train number character region is obtained, the segmentation image is subjected to HSL conversion through the HSL conversion module, the brightness of the image is obtained, the higher the brightness value is, the whiter the color is, the darker the color is, the color different from the black train number in the segmentation image is removed according to the brightness value, the black train number character image is left and is sent to the character recognition module for recognition, the automation degree is high, and the difficulty of train number recognition is reduced, the train number of the train can be effectively identified, the accuracy of train number identification is improved, the time of train number identification is shortened, and the train maintenance efficiency is improved.
In a third aspect, the present application provides a computer-readable storage medium storing a computer program, which adopts the following technical solutions:
a computer-readable storage medium storing a computer program, the computer program implementing the method of any one of the first aspects.
By adopting the technical scheme, the video data stream of the train collected by monitoring is obtained, the area of the movement change of the train is obtained according to the interframe information to obtain the image of the train, the image is preprocessed to remove noise interference, the definition and the quality of the image are improved, the image without the noise interference is positioned in the character area of the train number and is intercepted to obtain the segmentation image containing the character area of the train number, the segmentation image is subjected to HSL conversion to obtain the brightness of the image, the higher the brightness value is, the whiter the color is, the lower the brightness value is, the blacker the color is, the color different from the black train number in the segmentation image is removed according to the brightness value, the black character image of the train number is left to be identified, the automation degree is high, the difficulty of identifying the train number is reduced, the train number of the train can be effectively identified, the accuracy of identifying the train number of the train is improved, and the time of identifying the train number is reduced, the efficiency of train maintenance is improved.
In summary, the invention has the following beneficial technical effects:
1. according to the method, a monitored and collected train video data stream is obtained, firstly, a train movement change area is obtained according to interframe information to obtain an image of a train, the image is preprocessed to remove noise interference, a preset character area positioning model is used for carrying out train number character area positioning and intercepting on the image without the noise interference to obtain a segmentation image containing a train number character area, the segmentation image is subjected to HSL conversion, the segmentation image is processed by utilizing L brightness information to obtain a black train number character image, and the black train number character image is identified through a preset light-weight CRNN network model, so that the train number can be accurately identified;
2. the invention has high automation degree, reduces the difficulty of train number identification, can effectively identify the train number of the train, improves the accuracy of train number identification, reduces the time of train number identification and improves the efficiency of train maintenance.
Drawings
FIG. 1 is a flow chart of a method for identifying a train number in a colored drawing according to an embodiment of the present invention;
fig. 2 is a block diagram illustrating a method for identifying a train number by color drawing according to an embodiment of the present invention;
fig. 3 is a diagram of a lightweight CRNN network model.
Description of reference numerals:
1. a video acquisition module;
2. a noise removal module;
3. a character positioning module;
4. an HSL transform module;
5. and a character recognition module.
Detailed Description
The present invention is described in further detail below with reference to fig. 1-3.
The embodiment of the invention discloses a colored drawing train number identification method.
In an alternative embodiment, the identification method is shown in fig. 1, and includes the following specific steps:
s1, acquiring a train video data stream collected by monitoring, acquiring interframe information through the video data stream, and acquiring an image of a train motion change area according to the interframe information;
and the dynamic image is obtained by utilizing the correlation of the information between the front frame and the rear frame, so that the character area of the image is conveniently positioned and identified.
S2, preprocessing the image and removing noise interference;
and removing noise interference by adopting a filtering algorithm, specifically, removing noise interference by adopting Gaussian filtering, wherein the Gaussian filtering is a process of carrying out weighted average on the whole image, and the value of each pixel point is obtained by carrying out weighted average on the value of each pixel point and other pixel values in the neighborhood.
After the noise interference is removed from the image, the definition and the quality of the image can be improved, and the subsequent work such as image character region positioning, recognition and the like is facilitated.
S3, positioning the car number character region of the image without noise interference to obtain a segmentation image containing the car number character region;
by segmenting the image of the positioned vehicle number character region, the influence of the non-vehicle number character region on character recognition is reduced, and the accuracy of vehicle number recognition is improved.
Wherein, step S3 specifically includes:
acquiring an image without noise interference, and sending the image to a preset character area positioning model;
performing car number character area positioning on the image without noise interference by using the preset character area positioning model;
and segmenting the image of the positioned car number character area to obtain a segmented image containing the car number character area.
The specific generation steps of the character area positioning model are as follows:
acquiring an initial model and a first data set, wherein the first data set comprises a conventional train picture and a colored drawing train picture;
carrying out region marking on characters in a train number picture of a train to form a first training set and a first testing set;
performing target detection network training on the initial model according to the first training set and the first test set to generate a character area positioning model;
the method comprises the steps of carrying out region labeling on characters in a train number picture of a train, wherein the regions refer to positions of the characters in the picture, the region labeling refers to labeling of the positions of the characters through region labeling software, and the region labeling software can be labelme.
It should be noted that the training process of the target detection network is divided into two stages, and the first stage is a stage in which data is propagated from a low level to a high level, i.e. a forward propagation stage. The other stage is a stage for training the propagation of the error from the high level to the low level when the result of the current propagation does not match the expectation, i.e. a back propagation stage. The specific process is as follows:
1. initializing a weight value by the network;
2. input data is transmitted forwards through a network to obtain an output value;
3. calculating the error between the output value of the network and the target value;
4. when the error is larger than our expected value, the error is transmitted back to the network, and the updated weight value is propagated backwards.
5. And repeating the steps 2-4 until the error is equal to or less than the expected value, and finishing the training.
S4, performing HSL conversion processing on the divided images;
the image is divided into RGB pictures according to the divided image including the car number character area, the divided image is HSL converted, and the divided image is processed according to L information in the HSL conversion.
By performing the HSL conversion on the image, the brightness L of the image, i.e., the brightness of the color, can be obtained, the higher the brightness value, the whiter the color, the lower the brightness value, and the blacker the color, and the color different from the black car number in the divided image can be removed according to the obtained brightness value, leaving the black car number character information.
The specific treatment process is as follows:
let RGB be (L) R ，L G ，L B ) Is provided with L max Is L R ，L G ，L B Maximum value of, L min Is L R ，L G ，L B Minimum value of, L R ，L G ，L B Has a value interval of [0,1 ]]；
(1) Calculation formula of luminance L: l = (L) max +L min ）/2
In a special case, when L max =L min At this time, there is L max =L R =L G =L B =L min Indicating the color as grey, when S =0, H does not represent any color;
(2) the calculation of the saturation S is divided into two cases:
if L is less than or equal to 0.5, S = (L) max -L min ）/（L max +L min ）；
If L > 0.5, S = (L) max -L min ）/（2-L max -L min ）；
(3) The formula for the hue H is calculated as follows in three cases:
when L is max =L R H =60 (L) G -L B ）/（L max +L min ) The color is between yellow and magenta;
when L is max =L G H =120+60 (L) B -L R ）/（L max +L min ) The color is between cyan and yellow;
when L is max =L B H =240+60 (L) R -L G ）/（L max +L min ) The color is between magenta and cyan;
if the above calculation H takes on a negative value, it is increased 360, since H is a periodic function.
Wherein HSL is Hue, Saturation, brightness (English: Hue, Saturation, Lightness),
hue (H) is a basic attribute of color, which is a commonly-known color name, such as red, yellow, etc.; the saturation (S) is the purity of the color, the higher the color is, the more pure the color is, the lower the color is, the gray gradually becomes, and the numerical value of 0-100% is taken; the brightness (L) refers to the brightness of the color, the higher the brightness value is, the whiter the color is, the lower the brightness is, the blacker the color is; lightness (V) and brightness (L), 0-100%;
after HSL transformation, counting the brightness value of each pixel point, wherein the higher the brightness value is, the whiter the color is, the lower the brightness value is, the blacker the color is, selecting a proper threshold value as a critical point, establishing an image template (the point with the brightness value larger than the threshold value, the value of the pixel point in the image template is 255, otherwise 0), and then matching the image template with the original RGB image, so that only the black car number information is stored.
S5, performing character recognition on the image subjected to the HSL conversion processing;
step S5 specifically includes:
acquiring an image subjected to HSL conversion processing, and sending the image to a pre-generated lightweight CRNN network model;
recognizing characters in the image subjected to the HSL conversion processing by utilizing a pre-generated lightweight CRNN network model;
the specific generation steps of the lightweight CRNN network model are as follows:
acquiring a second data set, wherein the second data set is a conventional train picture and a colored drawing train picture;
processing the second data set through a character area positioning model to obtain a segmentation image containing a car number character area, performing HSL (hue, saturation and lightness) conversion on the segmentation image, and processing the segmentation image by utilizing brightness L information to obtain an image to form a second training set and a second test set;
performing network training according to the second training set and the second testing set to generate a lightweight CRNN network model;
carrying out lightweight to the character recognition model, obtaining a lightweight CRNN network model, wherein the lightweight CRNN network model comprises three parts, which are sequentially from bottom to top:
(1) and (3) rolling layers: extracting a characteristic sequence from an input image;
(2) circulating layer: predicting a label distribution of the feature sequence obtained from the convolutional layer;
(3) transcription layer: and converting the label distribution acquired from the circulation layer into a final identification result through operations of de-duplication, integration and the like.
Fig. 3 is a diagram of a lightweight CRNN network model.
The lightweight CRNN network model is small in size after being lightened, and the character recognition time is shortened.
Based on the colored drawing train number identification method, the embodiment of the application also discloses a colored drawing train number identification system.
In an optional embodiment, the method specifically includes:
the video acquisition module 1 is used for acquiring video data of train running, acquiring interframe information through the video data and acquiring an image of a train motion change area according to the interframe information;
the noise removal module 2 is used for preprocessing the acquired images of the train motion change area and removing noise interference;
the character positioning module 3 is used for positioning the car number character area of the image without the noise interference to obtain a segmented image containing the car number character area;
an HSL conversion module 4 for performing HSL conversion processing on the divided image including the car number character region;
and a character recognition module 5 for performing character recognition on the image subjected to the HSL conversion processing.
In the working process of the embodiment of the invention, video data of train running is obtained through the video acquisition module 1, interframe information is obtained from the video data, and an image of a train movement change area is obtained according to the interframe information; the method comprises the steps of sending an image of a train movement change area to a noise removal module 2 to remove noise interference, sending the image after the noise interference is removed to a character positioning module 3 to position the image in a train number character area, intercepting according to the positioned train number character area position to obtain a segmentation image containing the train number character area, sending the segmentation image to an HSL conversion module 4 to carry out HSL conversion, processing the segmentation image according to L information in the HSL conversion, removing colors different from the train number in the segmentation image, leaving the train number character image, and sending the segmentation image to a character recognition module 5 to recognize train number characters.
The embodiment of the application also discloses a computer readable storage medium storing a computer program, and the computer program for realizing any one of the methods is stored.
The above are all preferred embodiments of the present invention, and the protection scope of the present invention is not limited thereby, so: all equivalent changes made according to the structure, shape and principle of the invention are covered by the protection scope of the invention.
Claims (9)
1. A colored drawing train number identification method is characterized by comprising the following steps:
acquiring a train video data stream collected by monitoring, acquiring interframe information through the video data stream, and acquiring an image of a train motion change area according to the interframe information;
preprocessing an image to remove noise interference;
positioning the car number character region of the image without noise interference to obtain a segmented image containing the car number character region;
performing HSL transformation processing on the segmentation image;
and performing character recognition on the image subjected to the HSL conversion processing.
2. The colored drawing train number identification method according to claim 1, characterized in that: the image is preprocessed, and the noise interference is removed specifically as follows:
and removing noise interference by Gaussian filtering by adopting a filtering algorithm.
3. The colored drawing train number identification method according to claim 2, wherein the image without noise interference is subjected to train number character area positioning to obtain a segmented image containing a train number character area, and the steps specifically comprise:
acquiring an image without noise interference, and sending the image to a preset character area positioning model;
performing car number character area positioning on the image without noise interference by using the preset character area positioning model;
and segmenting the image of the positioned car number character area to obtain a segmented image containing the car number character area.
4. The colored drawing train number identification method according to claim 3, wherein the specific generation steps of the character area positioning model are as follows:
acquiring an initial model and a first data set, wherein the first data set comprises a conventional train picture and a colored drawing train picture;
carrying out region labeling on characters in the train number picture of the train to form a first training set and a first testing set;
and performing target detection network training on the initial model according to the first training set and the first testing set to generate a character area positioning model.
5. The colored drawing train number identification method according to claim 4, wherein the HSL conversion processing is performed on the segmentation image, and the steps specifically comprise:
the method comprises the steps of performing HSL conversion on a divided image containing a car number character area according to the divided image, and processing the divided image according to L information in the HSL conversion.
6. The colored drawing train number recognition method according to claim 5, wherein the step of performing character recognition on the image subjected to HSL conversion specifically comprises the steps of:
acquiring an image subjected to HSL conversion processing, and sending the image to a pre-generated lightweight CRNN network model;
and recognizing characters in the image subjected to the HSL conversion processing by using the pre-generated lightweight CRNN network model.
7. The colored drawing train number identification method as claimed in claim 6, wherein the specific generation steps of the light-weighted CRNN network model are as follows:
acquiring a second data set, wherein the second data set is a conventional train picture and a colored drawing train picture;
processing the second data set through the character region positioning model to obtain a car number character region positioning segmentation picture, and processing the segmentation picture by using HSL (hue, saturation and value) conversion information of the segmentation picture to form a second training set and a second test set;
and carrying out network training according to the second training set and the second testing set to generate a lightweight CRNN network model.
8. The utility model provides a system for colored drawing train number discernment which characterized in that includes:
the video acquisition module (1) is used for acquiring video data of train running, acquiring interframe information through the video data and acquiring an image of a train movement change area according to the interframe information;
the noise removing module (2) is used for preprocessing the acquired images of the train motion change area and removing noise interference;
the character positioning module (3) is used for positioning the car number character region of the image without the noise interference to obtain a segmented image containing the car number character region;
an HSL conversion module (4) for performing HSL conversion processing on the divided image including the car number character region;
and a character recognition module (5) for performing character recognition on the image subjected to the HSL conversion processing.
9. A computer-readable storage medium storing a computer program, characterized in that a computer program implementing the method of any one of claims 1 to 7 is stored.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210744590.3A CN114821452B (en) | 2022-06-29 | 2022-06-29 | Colored drawing train number identification method, system and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210744590.3A CN114821452B (en) | 2022-06-29 | 2022-06-29 | Colored drawing train number identification method, system and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114821452A true CN114821452A (en) | 2022-07-29 |
| CN114821452B CN114821452B (en) | 2022-09-20 |
Family
ID=82523392
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210744590.3A Active CN114821452B (en) | 2022-06-29 | 2022-06-29 | Colored drawing train number identification method, system and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114821452B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117173416A (en) * | 2023-11-01 | 2023-12-05 | 山西阳光三极科技股份有限公司 | Railway freight train number image definition processing method based on image processing |
| CN117315664A (en) * | 2023-09-18 | 2023-12-29 | 山东博昂信息科技有限公司 | Scrap steel bucket number identification method based on image sequence |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101398894A (en) * | 2008-06-17 | 2009-04-01 | 浙江师范大学 | Automobile license plate automatic recognition method and implementing device thereof |
| CN105354574A (en) * | 2015-12-04 | 2016-02-24 | 山东博昂信息科技有限公司 | Vehicle number recognition method and device |
| US20170024619A1 (en) * | 2015-07-22 | 2017-01-26 | Xerox Corporation | Video-based system and method for parking occupancy detection |
| CN106815580A (en) * | 2016-12-23 | 2017-06-09 | 上海集成电路研发中心有限公司 | A kind of license plate locating method |
-
2022
- 2022-06-29 CN CN202210744590.3A patent/CN114821452B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101398894A (en) * | 2008-06-17 | 2009-04-01 | 浙江师范大学 | Automobile license plate automatic recognition method and implementing device thereof |
| US20170024619A1 (en) * | 2015-07-22 | 2017-01-26 | Xerox Corporation | Video-based system and method for parking occupancy detection |
| CN105354574A (en) * | 2015-12-04 | 2016-02-24 | 山东博昂信息科技有限公司 | Vehicle number recognition method and device |
| CN106815580A (en) * | 2016-12-23 | 2017-06-09 | 上海集成电路研发中心有限公司 | A kind of license plate locating method |
Non-Patent Citations (1)
| Title |
|---|
| 华春梦 等: "《一种基于CRNN 的车牌识别算法研究与应用》", 《现代信息科技》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117315664A (en) * | 2023-09-18 | 2023-12-29 | 山东博昂信息科技有限公司 | Scrap steel bucket number identification method based on image sequence |
| CN117315664B (en) * | 2023-09-18 | 2024-04-02 | 山东博昂信息科技有限公司 | Scrap steel bucket number identification method based on image sequence |
| CN117173416A (en) * | 2023-11-01 | 2023-12-05 | 山西阳光三极科技股份有限公司 | Railway freight train number image definition processing method based on image processing |
| CN117173416B (en) * | 2023-11-01 | 2024-01-05 | 山西阳光三极科技股份有限公司 | Railway freight train number image definition processing method based on image processing |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114821452B (en) | 2022-09-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114821452B (en) | Colored drawing train number identification method, system and medium | |
| CN109872285B (en) | Retinex low-illumination color image enhancement method based on variational constraint | |
| CN109145915A (en) | License plate rapid distortion antidote under a kind of complex scene | |
| CN112287941B (en) | License plate recognition method based on automatic character region perception | |
| CN103119625B (en) | Video character separation method and device | |
| CN111709888B (en) | Aerial image defogging method based on improved generation countermeasure network | |
| CN111008608B (en) | Night vehicle detection method based on deep learning | |
| CN112561899A (en) | Electric power inspection image identification method | |
| CN113191216B (en) | Multi-user real-time action recognition method and system based on posture recognition and C3D network | |
| CN112613579A (en) | Model training method and evaluation method for human face or human head image quality and selection method for high-quality image | |
| CN116309607B (en) | Ship type intelligent water rescue platform based on machine vision | |
| CN107360344A (en) | Monitor video rapid defogging method | |
| CN111815528A (en) | Bad weather image classification enhancement method based on convolution model and feature fusion | |
| CN112070691B (en) | Image defogging method based on U-Net | |
| CN112487981A (en) | MA-YOLO dynamic gesture rapid recognition method based on two-way segmentation | |
| US7620246B2 (en) | Method and apparatus for image processing | |
| CN117173416B (en) | Railway freight train number image definition processing method based on image processing | |
| CN112232205B (en) | Mobile terminal CPU real-time multifunctional face detection method | |
| CN116343100B (en) | Target identification method and system based on self-supervision learning | |
| CN111882495B (en) | Image highlight processing method based on user-defined fuzzy logic and GAN | |
| CN112381088A (en) | License plate recognition method and system for oil tank truck | |
| CN112907469B (en) | Underwater image identification method based on Lab domain enhancement, classification and contrast improvement | |
| CN111627047A (en) | Underwater fish dynamic visual sequence moving target detection method | |
| CN112116580B (en) | Detection method, system and equipment for camera support | |
| EP4276736A1 (en) | Using integrated system to enhance a reference image |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |