CN111797832A - Automatic generation method and system of image interesting region and image processing method - Google Patents
Automatic generation method and system of image interesting region and image processing method Download PDFInfo
- Publication number
- CN111797832A CN111797832A CN202010673144.9A CN202010673144A CN111797832A CN 111797832 A CN111797832 A CN 111797832A CN 202010673144 A CN202010673144 A CN 202010673144A CN 111797832 A CN111797832 A CN 111797832A
- Authority
- CN
- China
- Prior art keywords
- preset
- region
- image
- target
- interest
- 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
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000003672 processing method Methods 0.000 title claims abstract description 8
- 238000012549 training Methods 0.000 claims abstract description 37
- 238000012545 processing Methods 0.000 claims abstract description 22
- 238000010586 diagram Methods 0.000 claims abstract description 10
- 238000001914 filtration Methods 0.000 claims description 15
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000003708 edge detection Methods 0.000 claims description 3
- 230000001131 transforming effect Effects 0.000 claims 1
- 238000012986 modification Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 5
- 238000002372 labelling Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 230000002093 peripheral effect Effects 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
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an automatic generation method and system of an image interesting region and an image processing method, which relate to the field of image processing and comprise the following steps: acquiring a first image shot in a preset scene, and marking a preset target in the first image to obtain training data; collecting position information of a preset target in training data; obtaining a first position diagram of a preset target in a preset scene based on the position information of the preset target; training a target recognition network by using the training data to obtain a trained target recognition network; acquiring a second image shot in a preset scene, and recognizing a preset target in the second image by using a trained target recognition network to obtain second position maps of a plurality of preset targets; the method and the device can automatically calibrate the interested region in the image, improve the working efficiency of computer vision software development and save the labor and economic cost.
Description
Technical Field
The invention relates to the field of video image processing, in particular to an automatic generation method and system of an image interesting region and an image processing method.
Background
In the field of image processing, a region of interest (ROI) is an image region selected from an image, which is a region of major interest for image analysis. The area is delineated for further processing. The ROI is used for delineating the target which the user wants to read, so that the processing time can be reduced, and the precision can be increased. In the traditional region of interest division, a range needs to be manually marked on an image in advance, and different scenes need to be marked again.
In the application of big data, developers can use the data after marking extremely large-scale data, and the data marking usually needs manual marking, which has great time and money cost consumption.
If the usage amount of the marking data is directly reduced, the accuracy of the region of interest is directly reduced, and the problem of missing marking occurs. If a large amount of labeling data is used to improve the accuracy of defining the region of interest, a huge cost problem will be caused.
Disclosure of Invention
The invention provides an automatic generation method and system of an image region of interest and an image processing method, which are used for automatically calibrating the region of interest in an image. The method and the system can avoid the trouble that the interested region in the image is manually defined again when different scenes are defined, and improve the working efficiency of computer vision software development. The method and the system reduce the requirement of marking data and save labor and economic cost.
In order to achieve the above object, the present invention provides an automatic generation method of an image region of interest, including:
collecting a plurality of first images shot in a preset scene, and marking preset targets in the first images to obtain training data;
collecting position information of a preset target in training data;
acquiring a first position diagram of a preset target in a preset scene based on the acquired position information of the preset target;
training a target recognition network by using the training data to obtain a trained target recognition network;
acquiring a plurality of second images shot in a preset scene, and recognizing preset targets in the second images by using a trained target recognition network to obtain second position maps of a plurality of preset targets;
and fusing the first position map and the plurality of second position maps to generate the region of interest.
The method comprises the following steps: the method is based on semi-supervised learning and big data theory, identifies the target in the scene through a pre-training network, records the position of the identified target, and automatically generates the interested area of the scene by utilizing Gaussian kernel convolution and related image processing technology. The trouble that the region of interest needs to be manually calibrated again when the region of interest in the images is calibrated in different scenes can be avoided, the working efficiency of computer vision software development can be improved, the requirement for marking data is reduced, and the labor and economic cost is saved. Meanwhile, the usable data volume can be remarkably increased, and therefore the definition precision of the region of interest is improved. The method and the system reduce the engineering difficulty of using the region of interest, and enable the region of interest processing to be more widely used in the engineering practice in the computer vision field, thereby improving the accuracy and efficiency of computer image processing.
Preferably, the number of second images is greater than the number of first images. That is, only a small amount of marked data is needed to complete the processing of a large amount of unmarked data.
Preferably, the method further comprises:
clutter filtering to generate an interested area;
and binarizing the region of interest after clutter filtering.
The clutter filtering processing can effectively eliminate spots in the image, namely, remove some peripheral noise points, improve the definition precision of the region of interest, and then segment the scene into a target region where the target is likely to appear and a background region where the target is unlikely to appear.
Preferably, the clutter filtering generates the region of interest, specifically: and carrying out edge detection on the image, and eliminating the spots with the detected side length less than 5 pixels.
Preferably, the region of interest after the binarization clutter filtering specifically includes: setting the pixels larger than 1 in the image as 1, and setting the rest pixels as 0.
Preferably, the obtaining a first position map of the preset target in the preset scene based on the position information of the preset target specifically includes: based on the position information of the preset target, converting the preset target into a circular spot covering the preset target by utilizing Gaussian kernel convolution, and obtaining a first position diagram of the preset target in a preset scene.
Preferably, the formula for converting the preset target into a circular spot covering the preset target by using gaussian kernel convolution is as follows:
wherein G (x, y) is a single target appearance position diagram, x is an abscissa, y is an ordinate, pi is a circumference ratio, sigma is a Gaussian kernel size, and e is a natural constant.
Preferably, the formula for generating the region of interest by fusing the first location map and the plurality of second location maps is as follows:
ROI=∑mG(x,y)
wherein, ROI is the interested region, m is the picture number, and G (x, y) is the single target appearance position map. The first position diagram and the plurality of second position diagrams are fused to generate the region of interest, so that the advantages of big data can be fully utilized, the probability of the position of the target is close to the real situation statistically as much as possible, and the coverage area of the region of interest obtained by the method is more comprehensive and accurate.
Meanwhile, the invention also provides an automatic generation system of the image region of interest, which comprises the following steps:
the training data acquisition unit is used for acquiring a plurality of first images shot in a preset scene, and marking preset targets in the first images to obtain training data;
the preset target position information acquisition unit is used for acquiring position information of a preset target in the training data;
the first position map obtaining unit is used for obtaining a first position map of the preset target in a preset scene based on the collected position information of the preset target;
the training unit is used for training the target recognition network by using the training data to obtain the trained target recognition network;
the second position map obtaining unit is used for acquiring a plurality of second images shot in a preset scene, and recognizing preset targets in the second images by using the trained target recognition network to obtain second position maps of a plurality of preset targets;
and the interested region generating unit is used for fusing the first position map and the plurality of second position maps to generate the interested region.
Wherein, this system still includes:
the clutter filtering unit is used for filtering clutter to generate an interested region;
and the binarization unit is used for binarizing the region of interest after clutter filtering.
The invention also provides an image processing method, which comprises the following steps:
collecting an original image;
processing the original image by adopting the automatic generation method of the image interesting region to obtain the interesting region in the original image;
multiplying the region-of-interest image with the original image, removing a fixed background in the original image, and reserving a region where a target possibly appears to obtain a new image after the region-of-interest is processed;
inputting the new image after the region of interest processing into the neural network for processing, including but not limited to: and identifying, detecting and counting the target.
One or more technical schemes provided by the invention at least have the following technical effects or advantages:
the method and the system for automatically demarcating the region of interest based on semi-supervised learning and big data fusion are provided for solving the problems that the conventional region of interest needs to be manually demarcated, the region of interest needs to be demarcated again for different scenes, and time and labor are consumed. The method and the system can automatically define the region outside the background from the scene, solve the problem that software developers need to manually define the region of interest again when facing a new scene, remove the trouble of background interference, and effectively improve the working efficiency and the development progress of the software developers. Meanwhile, the method and the system use a semi-supervised learning mode, so that the demand of marking data can be reduced, the utilization rate of the data is improved, and the cost is saved.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the principles of the invention;
FIG. 1 is a flow chart of a method for automatically generating a region of interest in an image according to the present invention;
fig. 2 is a schematic diagram of the automatic generation system of the region of interest in the image according to the present invention.
Detailed Description
In order that the above objects, features and advantages of the present invention can be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings. It should be noted that the embodiments and features of the embodiments of the present application may be combined with each other without conflicting with each other.
In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced in other ways than those specifically described and thus the scope of the present invention is not limited by the specific embodiments disclosed below.
Example one
The embodiment of the invention provides an automatic generation method of an interested area in an image, which is based on semi-supervised learning and big data theory, identifies a target in a scene through a pre-training network, records the position of the identified target, and automatically generates the interested area of the scene by utilizing Gaussian kernel convolution and related image processing technology.
Referring to fig. 1, the automatic generation process of the region of interest is as follows:
collecting a plurality of images shot in a preset scene, and labeling preset targets in the images to obtain training data;
training a target recognition network by using the training data to obtain a trained target recognition network, wherein the target recognition network can recognize a preset target in the image, and the target recognition network can be a traditional VGG, Resnet, SSD or YOLO algorithm network, but also can be a self-defined algorithm network;
collecting the position of a target in the existing training data;
counting target positions in the existing training data; and converting the identified target in the step into a circular patch covering the target by utilizing Gaussian kernel convolution. This results in a first position map of the object that may occur in the predetermined scene. The mathematical principle is as follows:
wherein x is an abscissa, y is an ordinate, pi is a circumference ratio, sigma is a Gaussian kernel size, and e is a natural constant.
And expanding the target position data by using a semi-supervised learning principle. And carrying out target identification of the non-interesting area processing picture on the target scene by using a pre-trained network. And acquiring a large number of pictures after identifying the target to obtain a large number of second position pictures containing the target.
And generating a region of interest. And fusing the first position map and the generated second position maps of the targets into one map:
ROI=∑mG(x,y)
wherein, ROI is the interested region, m is the picture number, and G (x, y) is a single target appearance position map.
And performing clutter filtering on the region of interest. And (5) carrying out edge detection on the image, and eliminating the detected spots with the side length smaller than 5 pixels.
And carrying out binarization processing on the region of interest. Setting the pixel value less than 1 in the image as 0, and setting the pixel values of the rest pixels as 1. A region of interest map is obtained that can be used directly to cut the original image.
Where a is the pixel value.
Referring to fig. 2, a second embodiment of the present invention provides an automatic generation system of an image region of interest, including:
the training data acquisition unit is used for acquiring a plurality of first images shot in a preset scene, and marking preset targets in the first images to obtain training data;
the preset target position information acquisition unit is used for acquiring position information of a preset target in the training data;
the first position map obtaining unit is used for obtaining a first position map of a preset target in a preset scene based on the position information of the preset target;
the training unit is used for training the target recognition network by using the training data to obtain the trained target recognition network;
the second position map obtaining unit is used for acquiring a plurality of second images shot in a preset scene, and recognizing preset targets in the second images by using the trained target recognition network to obtain second position maps of a plurality of preset targets;
and the interested region generating unit is used for fusing the first position map and the plurality of second position maps to generate the interested region.
Wherein, this system still includes:
the clutter filtering unit is used for filtering clutter to generate an interested region;
and the binarization unit is used for binarizing the region of interest after clutter filtering.
EXAMPLE III
The third embodiment of the invention provides an image processing method, which comprises the following steps:
collecting an original image;
processing the original image by adopting the automatic generation method of the image interesting region to obtain the interesting region in the original image;
multiplying the region-of-interest image with the original image, removing a fixed background in the original image, and reserving a region where a target possibly appears to obtain a new image after the region-of-interest is processed;
inputting the new image after the region of interest processing into the neural network for processing, including but not limited to: and identifying, detecting and counting the target.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. Therefore, it is intended that the appended claims be interpreted as including preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.
Claims (10)
1. An automatic generation method of an image region of interest, characterized in that the method comprises:
collecting a plurality of first images shot in a preset scene, and marking preset targets in the first images to obtain training data;
collecting position information of a preset target in training data;
acquiring a first position diagram of a preset target in a preset scene based on the acquired position information of the preset target;
training a target recognition network by using the training data to obtain a trained target recognition network;
acquiring a plurality of second images shot in a preset scene, and recognizing preset targets in the second images by using a trained target recognition network to obtain second position maps of a plurality of preset targets;
and fusing the first position map and the plurality of second position maps to generate the region of interest.
2. The method of claim 1, wherein the number of the second images is greater than the number of the first images.
3. The method for automatically generating an image region of interest according to claim 1, further comprising:
clutter filtering to generate an interested area;
and binarizing the region of interest after clutter filtering.
4. The method for automatically generating an image region of interest according to claim 3, wherein the clutter filtering generated region of interest is specifically: and carrying out edge detection on the image, and eliminating the spots with the detected side length less than 5 pixels.
5. The method for automatically generating an image region of interest according to claim 3, wherein the binarized clutter filtered region of interest specifically includes: setting the pixels larger than 1 in the image as 1, and setting the rest pixels as 0.
6. The method for automatically generating an image region of interest according to claim 1, wherein the obtaining a first position map of a preset target in a preset scene based on position information of the preset target specifically includes: based on the position information of the preset target, converting the preset target into a circular spot covering the preset target by utilizing Gaussian kernel convolution, and obtaining a first position diagram of the preset target in a preset scene.
7. The method for automatically generating an image region of interest according to claim 6, wherein the formula for transforming the preset target into a circular spot covering the preset target by using the Gaussian kernel convolution is as follows:
wherein G (x, y) is a single target appearance position diagram, x is an abscissa, y is an ordinate, pi is a circumference ratio, sigma is a Gaussian kernel size, and e is a natural constant.
8. The method for automatically generating an image region of interest according to claim 1, wherein the formula for generating the region of interest by fusing the first location map and the plurality of second location maps is as follows:
ROI=∑mG(x,y)
wherein, ROI is the interested region, m is the picture number, and G (x, y) is the single target appearance position map.
9. An automatic image region-of-interest generation system, the system comprising:
the training data acquisition unit is used for acquiring a plurality of first images shot in a preset scene, and marking preset targets in the first images to obtain training data;
the preset target position information acquisition unit is used for acquiring position information of a preset target in the training data;
the first position map obtaining unit is used for obtaining a first position map of the preset target in a preset scene based on the collected position information of the preset target;
the training unit is used for training the target recognition network by using the training data to obtain the trained target recognition network;
the second position map obtaining unit is used for acquiring a plurality of second images shot in a preset scene, and recognizing preset targets in the second images by using the trained target recognition network to obtain second position maps of a plurality of preset targets;
and the interested region generating unit is used for fusing the first position map and the plurality of second position maps to generate the interested region.
10. An image processing method, characterized in that the method comprises:
collecting an original image;
processing the original image by adopting the image interesting region automatic generation method of any one of claims 1 to 8 to obtain an interesting region in the original image;
multiplying the region-of-interest image with the original image, removing a fixed background in the original image, and reserving a region where a target possibly appears to obtain a new image after the region-of-interest is processed;
inputting the new image after the region of interest processing into the neural network for processing, including but not limited to: and identifying, detecting and counting the target.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010673144.9A CN111797832B (en) | 2020-07-14 | 2020-07-14 | Automatic generation method and system for image region of interest and image processing method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010673144.9A CN111797832B (en) | 2020-07-14 | 2020-07-14 | Automatic generation method and system for image region of interest and image processing method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111797832A true CN111797832A (en) | 2020-10-20 |
CN111797832B CN111797832B (en) | 2024-02-02 |
Family
ID=72806812
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010673144.9A Active CN111797832B (en) | 2020-07-14 | 2020-07-14 | Automatic generation method and system for image region of interest and image processing method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111797832B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113435447A (en) * | 2021-07-26 | 2021-09-24 | 杭州海康威视数字技术股份有限公司 | Image annotation method, device and system |
CN113744126A (en) * | 2021-08-06 | 2021-12-03 | Oppo广东移动通信有限公司 | Image processing method and device, computer readable medium and electronic device |
Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103177458A (en) * | 2013-04-17 | 2013-06-26 | 北京师范大学 | Frequency-domain-analysis-based method for detecting region-of-interest of visible light remote sensing image |
CN104573669A (en) * | 2015-01-27 | 2015-04-29 | 中国科学院自动化研究所 | Image object detection method |
US20150206004A1 (en) * | 2014-01-20 | 2015-07-23 | Ricoh Company, Ltd. | Object tracking method and device |
CN104835175A (en) * | 2015-05-26 | 2015-08-12 | 西南科技大学 | Visual attention mechanism-based method for detecting target in nuclear environment |
CN107590262A (en) * | 2017-09-21 | 2018-01-16 | 黄国华 | The semi-supervised learning method of big data analysis |
CN108629378A (en) * | 2018-05-10 | 2018-10-09 | 上海鹰瞳医疗科技有限公司 | Image-recognizing method and equipment |
CN109961834A (en) * | 2019-03-22 | 2019-07-02 | 上海联影医疗科技有限公司 | The generation method and equipment of diagnostic imaging report |
CN110503047A (en) * | 2019-08-26 | 2019-11-26 | 西南交通大学 | A kind of rds data processing method and processing device based on machine learning |
CN111091127A (en) * | 2019-12-16 | 2020-05-01 | 腾讯科技(深圳)有限公司 | Image detection method, network model training method and related device |
CN111161227A (en) * | 2019-12-20 | 2020-05-15 | 成都数之联科技有限公司 | Target positioning method and system based on deep neural network |
CN111368658A (en) * | 2020-02-24 | 2020-07-03 | 交通运输部水运科学研究所 | Automatic detection method and system for external target of intelligent ship in autonomous navigation |
-
2020
- 2020-07-14 CN CN202010673144.9A patent/CN111797832B/en active Active
Patent Citations (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103177458A (en) * | 2013-04-17 | 2013-06-26 | 北京师范大学 | Frequency-domain-analysis-based method for detecting region-of-interest of visible light remote sensing image |
US20150206004A1 (en) * | 2014-01-20 | 2015-07-23 | Ricoh Company, Ltd. | Object tracking method and device |
CN104573669A (en) * | 2015-01-27 | 2015-04-29 | 中国科学院自动化研究所 | Image object detection method |
CN104835175A (en) * | 2015-05-26 | 2015-08-12 | 西南科技大学 | Visual attention mechanism-based method for detecting target in nuclear environment |
CN107590262A (en) * | 2017-09-21 | 2018-01-16 | 黄国华 | The semi-supervised learning method of big data analysis |
CN108629378A (en) * | 2018-05-10 | 2018-10-09 | 上海鹰瞳医疗科技有限公司 | Image-recognizing method and equipment |
CN109961834A (en) * | 2019-03-22 | 2019-07-02 | 上海联影医疗科技有限公司 | The generation method and equipment of diagnostic imaging report |
CN110503047A (en) * | 2019-08-26 | 2019-11-26 | 西南交通大学 | A kind of rds data processing method and processing device based on machine learning |
CN111091127A (en) * | 2019-12-16 | 2020-05-01 | 腾讯科技(深圳)有限公司 | Image detection method, network model training method and related device |
CN111161227A (en) * | 2019-12-20 | 2020-05-15 | 成都数之联科技有限公司 | Target positioning method and system based on deep neural network |
CN111368658A (en) * | 2020-02-24 | 2020-07-03 | 交通运输部水运科学研究所 | Automatic detection method and system for external target of intelligent ship in autonomous navigation |
Non-Patent Citations (1)
Title |
---|
董志鹏等: "遥感影像目标的尺寸特征卷积神经网络识别法", 《测绘学报》, vol. 48, no. 10, pages 1285 - 1295 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113435447A (en) * | 2021-07-26 | 2021-09-24 | 杭州海康威视数字技术股份有限公司 | Image annotation method, device and system |
CN113435447B (en) * | 2021-07-26 | 2023-08-04 | 杭州海康威视数字技术股份有限公司 | Image labeling method, device and image labeling system |
CN113744126A (en) * | 2021-08-06 | 2021-12-03 | Oppo广东移动通信有限公司 | Image processing method and device, computer readable medium and electronic device |
Also Published As
Publication number | Publication date |
---|---|
CN111797832B (en) | 2024-02-02 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN113160192B (en) | Visual sense-based snow pressing vehicle appearance defect detection method and device under complex background | |
CN111695486B (en) | High-precision direction signboard target extraction method based on point cloud | |
CN109413411B (en) | Black screen identification method and device of monitoring line and server | |
CN114677554A (en) | Statistical filtering infrared small target detection tracking method based on YOLOv5 and Deepsort | |
CN111369605B (en) | Infrared and visible light image registration method and system based on edge features | |
WO2013088175A1 (en) | Image processing method | |
CN110288612B (en) | Nameplate positioning and correcting method and device | |
CN113392856B (en) | Image forgery detection device and method | |
CN111695373B (en) | Zebra stripes positioning method, system, medium and equipment | |
CN111126393A (en) | Vehicle appearance refitting judgment method and device, computer equipment and storage medium | |
CN111797832B (en) | Automatic generation method and system for image region of interest and image processing method | |
CN114049499A (en) | Target object detection method, apparatus and storage medium for continuous contour | |
CN108961262B (en) | Bar code positioning method in complex scene | |
EP3553700A2 (en) | Remote determination of containers in geographical region | |
CN115908988B (en) | Defect detection model generation method, device, equipment and storage medium | |
CN114882204A (en) | Automatic ship name recognition method | |
CN114119695A (en) | Image annotation method and device and electronic equipment | |
CN114155285A (en) | Image registration method based on gray level histogram | |
CN110276260B (en) | Commodity detection method based on depth camera | |
CN115908774B (en) | Quality detection method and device for deformed materials based on machine vision | |
Kaimkhani et al. | UAV with Vision to Recognise Vehicle Number Plates | |
CN111027560B (en) | Text detection method and related device | |
CN114663681A (en) | Method for reading pointer type meter and related product | |
CN111626299A (en) | Outline-based digital character recognition method | |
Liu et al. | Identification of Damaged Building Regions from High-Resolution Images Using Superpixel-Based Gradient and Autocorrelation Analysis |
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 | ||
CB02 | Change of applicant information | ||
CB02 | Change of applicant information |
Address after: 610042 No. 270, floor 2, No. 8, Jinxiu street, Wuhou District, Chengdu, Sichuan Applicant after: Chengdu shuzhilian Technology Co.,Ltd. Address before: No.2, floor 4, building 1, Jule road crossing, Section 1, West 1st ring road, Wuhou District, Chengdu City, Sichuan Province 610041 Applicant before: CHENGDU SHUZHILIAN TECHNOLOGY Co.,Ltd. |
|
GR01 | Patent grant | ||
GR01 | Patent grant |