CN109190616B - Hot-rolled steel plate online visual tracking method based on feature recognition - Google Patents
Hot-rolled steel plate online visual tracking method based on feature recognition Download PDFInfo
- Publication number
- CN109190616B CN109190616B CN201810879301.4A CN201810879301A CN109190616B CN 109190616 B CN109190616 B CN 109190616B CN 201810879301 A CN201810879301 A CN 201810879301A CN 109190616 B CN109190616 B CN 109190616B
- Authority
- CN
- China
- Prior art keywords
- steel plate
- actual
- width
- area
- mask
- 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.)
- Active
Links
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
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration by the use of local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- 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
- G06V10/255—Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- 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
Abstract
The invention discloses a hot rolled steel plate online visual tracking method based on feature recognition, which accurately recognizes the maximum outer contour of a steel plate by utilizing feature adaptive thresholds of photo-thermal, polygonal, size, online quantity and the like of a hot rolled steel plate to finish the visual tracking of the steel plate in the movement process. Through the characteristics, the interference of water vapor, background objects and other light sources to the identification of the steel plate can be effectively reduced or eliminated. The invention tests various working conditions of the hot rolling area, and ensures the universality and the stronger robustness of the method. The invention can realize the position tracking of the hot-rolled area on-line steel plate, provides accurate reference for the process execution of the whole flow, and has the tracking position precision of about 0.1m within the range of 50m after the on-line actual measurement.
Description
Technical Field
The invention relates to the field of visual identification, in particular to a hot rolled steel plate online visual tracking method based on feature identification.
Background
The steel plate tracking is an important component of automatic systems such as a through heating furnace, a rolling mill, cooling, straightening and the like in the hot rolling production process. Traditional tracking logic and data mainly depend on-site detection instruments, sensors and the like, such as hot detection, cold detection, photoelectric switches, speed transmitters and the like. The detection instruments transmit the roller way speed and the trigger signal of a certain position node to the basic automation to calculate and simulate the position of the steel plate.
However, in the actual production process, the actual speed of the steel plate is not completely matched with the rotation speed of the roller way due to friction, collision, inertia and the like, and simultaneously, water vapor, heat radiation and the like can interfere with instrument signals such as cold/heat detection and the like. These all can cause the calculated position of steel sheet to have great error with true position, can disturb the model calculation of each system even, and then influence production rhythm and product quality.
The machine vision recognition technology is widely applied to the industrial field at present, and has the advantages of high processing speed, high detection precision, low cost, easiness in integration and the like.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides an efficient and accurate hot rolled steel plate online visual tracking method, which applies visual identification to online tracking detection of hot rolled steel plates, utilizes characteristic attributes of the hot rolled steel plates to assist edge detection, and improves the accuracy of contour detection. The position of the on-line steel plate is calculated quantitatively, the tracking precision is improved, the automatic control level of the whole-flow production is improved, and meanwhile, most of detection instruments can be replaced, and the operation and maintenance cost is greatly reduced.
The above purpose is realized by the following scheme:
a hot rolled steel plate online visual tracking method based on feature recognition comprises the following steps:
(1) continuously shooting the on-line steel plate in the hot rolling area, and converting a single-frame picture into a gray matrix;
(2) setting a mask according to a preset area, and extracting an interested area;
(3) performing perspective transformation, and projecting the inclined steel plate in the picture to another space to make the steel plate in the picture aligned;
the perspective transformation is carried out according to the following formula, wherein u and v are original picture coordinates, and transformed picture coordinates x and y are correspondingly obtained;
x＝x′/w′
y＝y′/w′
where w is 1, w' is the intermediate variable calculated, aiiIn order to transform the matrix, the matrix is,
representing a linear transformation, T2＝[a13 a23]TGenerating a perspective transformation, T3＝[a31 a32]For translation, aiiAll can pass through the known point pair { (u)j,vj),(xj,yj) And (6) obtaining.
Processing by using closed operation of morphological transformation after perspective, connecting breakpoints and closing small holes; the closed operation processing process is a process of expansion and corrosion, wherein the corrosion is to convolute the constructed kernel on the original image, extract the minimum value in the kernel coverage area and replace the anchor point position (default is the kernel central point), and the white area is reduced because the minimum value is selected; the expansion refers to selecting the maximum value of the coverage area to refer to the pixel at the anchor point position, and the white area expands; the closed operation can effectively eliminate small black holes, smooth the contour, connect narrow gaps, and fill holes smaller than the structural elements.
The kernel matrix is as follows,
(4) according to the actual number of the on-line steel plates, carrying out first self-adaptive threshold binarization to complete profile primary selection;
(5) and performing secondary self-adaptive threshold binarization according to the characteristics of the approximate quadrangle of the steel plates and the actual size of each steel plate to finish final outline determination and steel plate position calculation of the steel plates.
Further, the first adaptive threshold procedure in the step (4) is as follows:
firstly, carrying out binarization by using a larger initial threshold value to obtain a rough area of a steel plate and other noise areas; traversing the outlines of all the areas and performing descending arrangement according to the areas of the areas, if the number of the found outlines is less than the number of the actual steel plates, circularly reducing the threshold value according to the set step length of 5-20 until the number of the detected outlines is more than or equal to the number of the actual steel plates; and marking the front actual steel plate number contours sorted according to the area after the number of the contours is larger than or equal to the actual steel plate number.
Further, the second adaptive threshold procedure in the step (5) is as follows:
reordering the first recognized contours according to barycentric coordinates, and independently performing the following operations for each contour corresponding to the actual online steel plate sequence position:
and calculating the width of the projected outline according to the following formula according to the actual width of the steel plate. Because the steel plate is approximate to a quadrangle, the rectangular frame of the steel plate under the threshold value is obtained by operating the minimum outline covering rectangle;
in the formula, wtFor the theoretical width of the steel plate after perspective transformation, wmask_pFor the calculated width of the mask region after perspective, wmask_aTo perspective the actual width of the front mask region, wplate_aIs the actual width of the steel plate;
comparing the width of the minimum coating rectangle with the theoretical width of the steel plate after perspective transformation, if the width of the minimum coating rectangle is smaller than the theoretical width of the steel plate after perspective transformation, circularly reducing the threshold value according to the specified step length of 5-20 until the width of the minimum coating rectangle is larger than the theoretical width of the steel plate after perspective transformation; the process is coarse adjustment, the threshold value obtained at the end of the process is circularly increased by 1, until the width of the minimum coating rectangle is just less than or equal to the theoretical width of the steel plate after perspective transformation, and the process is fine adjustment;
repeating the process for each outline, namely the actual steel plate in sequence to obtain a self-adaptive threshold value and a steel plate rectangular outline meeting the actual steel plate width;
and calculating the actual position of the online steel plate according to the position of each rectangular outline and the actual size corresponding to the projected mask.
Further, after the step (5), the method further comprises: and (5) the rectangular outline of the steel plate is back projected and restored to the original frame picture.
The invention has the beneficial effects that: the invention utilizes the characteristics of the hot-rolled steel plate such as light and heat, polygon, size, online quantity and the like to carry out visual identification and tracking on the steel plate, can effectively reduce or eliminate the interference of water vapor, background objects and other light sources on the identification of the steel plate, and has the tracking position precision of about 0.1m within the range of 50m through field test.
Drawings
FIG. 1 is a flow diagram of an online visual tracking system;
FIG. 2 is an online tracking and actual mapping diagram of a plurality of steel plates in a certain steel mill.
Detailed Description
The steel plate on-line tracking method based on feature search proposed by the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.
As shown in fig. 1, the present embodiment uses opencv3 self-contained functions and autonomously developed classes and algorithms to complete visual recognition:
(1) and (3) arranging the high-speed industrial camera at a position 30-50m away from the hot rolling roller way area, and shooting the roller way surface.
(2) And transmitting the shot video to a computing system in the form of a picture, checking a frame picture to be processed, and if the number of channels is more than 1, performing gray level conversion by using 'cvtColor'.
(3) The 4 vertex coordinates of the mask area are determined, the mask is made by using "fillPoly", and the mask is loaded on the original image by using "add".
(4) And performing perspective transformation, and calculating 4 vertexes of the region before perspective (mask region) and 4 vertexes of the region after perspective by using 'getPerspectivetTransform'. Then, the transformed image can be obtained by using the "perspectiveTransform".
(5) Because there are breakpoints and holes after perspective transformation, it needs to use closed operation processing of morphological transformation, and the specifically used functions are "getstructural element" (obtain kernel) and "morphologyEx" (closed operation).
(6) The image closed after perspective can be obtained through the steps, and a binary image is obtained by utilizing a threshold function, wherein the threshold value is initially set to be 240. And then finding the outlines of the steel plates through 'findContours', sequencing the found outlines according to the areas by using 'sort', if the number of the found outlines is less than the number of the actual steel plates, gradually reducing the threshold value according to a set step length 5 until the number of the found outlines is more than or equal to the number of the actual steel plates, and then marking the number of the outlines of the first steel plates which are sequenced according to the areas.
(7) The labeled contours are sorted by barycentric coordinates, using the function "moments". Next, a threshold value and a final rectangular profile are calculated for each profile (corresponding to the steel plate) separately.
(8) Taking one contour sorted above, and obtaining a minimum coating rectangle by using 'boundingRef'. The theoretical width of the steel plate after perspective transformation is obtained according to the following formula.
In the formula, wtFor the theoretical width of the steel plate after perspective transformation, wmask_pFor the calculated width of the mask region after perspective, wmask_aTo perspective the actual width of the front mask region, wplate_aIs the actual width of the steel plate.
(9) If the width of the minimum coating rectangle is smaller than the theoretical width of the steel plate after perspective transformation, the threshold value is gradually reduced according to a set step length 5, then binarization, closed operation, contour finding, minimum coating rectangle finding and comparison of the two widths are carried out until the condition is met. This step is a coarse adjustment.
(10) And (4) repeating the calculation similar to the step (8), wherein the condition is changed to that the width of the minimum coating rectangle is less than or equal to the theoretical width of the steel plate after perspective transformation, and the step length of the threshold value is set to be 1, namely, the threshold value obtained in the step (8) is increased until the width of the minimum coating rectangle is less than or equal to the actual width of the steel plate. This step is fine tuning.
(11) And (5) repeating the steps (7), (8) and (9) for each contour, and finally outputting the found rectangular boxes and the corresponding threshold values.
(12) According to the head position p of each found rectangular framei(i is less than or equal to the number of the actual steel plates), and calculating the actual head position of each steel plate according to the following formula.
In the formula, wherein P _ aiIs the actual head position of the steel plate, piIs the head position of the rectangular frame, LmaskIs the length of the mask region after projection, LactFor real length of mask region, PmaskIs the distance of the starting position of the mask area from the target.
(13) And the rectangular frame back projection is restored to the original frame picture, and the steel plate position data calculated in real time is transmitted to a roller way transmission control system, so that the fine control of the full-flow material is realized.
Claims (3)
1. A hot rolled steel plate online visual tracking method based on feature recognition is characterized by comprising the following steps:
(1) continuously shooting the on-line steel plate in the hot rolling area, and converting a single-frame picture into a gray matrix;
(2) setting a mask according to a preset area, and extracting an interested area;
(3) performing perspective transformation, and projecting the inclined steel plate in the picture to another space to make the steel plate in the picture aligned;
the perspective transformation is carried out according to the following formula, wherein u and v are original picture coordinates, and transformed picture coordinates x and y are correspondingly obtained;
x＝x′/w′
y＝y′/w′
where w is 1, w' is the intermediate variable calculated, aiiIn order to transform the matrix, the matrix is,
representing a linear transformation, T2＝[a13 a23]TGenerating a perspective transformation, T3＝[a31 a32]For translation, aiiAll can pass through the known point pair { (u)j,vj),(xj,yj) Obtaining;
processing by using closed operation of morphological transformation after perspective, connecting breakpoints and closing small holes; the closed operation processing process is a process of expansion and corrosion, wherein the corrosion is to convolute the constructed kernel on the original image, extract the minimum value in the kernel coverage area to replace the anchor point position, and reduce the white area; the expansion refers to selecting the maximum value of the coverage area to refer to the pixel at the anchor point position, and the white area expands;
(4) according to the actual number of the on-line steel plates, carrying out first self-adaptive threshold binarization to complete profile primary selection;
(5) according to the characteristics of the approximate quadrangle of the steel plates and the actual size of each steel plate, performing secondary self-adaptive threshold binarization to complete final outline determination and steel plate position calculation of the steel plates; the second adaptive threshold process is as follows:
reordering the first recognized contours according to barycentric coordinates, and independently performing the following operations for each contour corresponding to the actual online steel plate sequence position:
calculating the width of the projected outline according to the actual width of the steel plate according to the following formula; performing operation on the minimum outline coated rectangle to obtain a rectangular frame of the steel plate under the threshold value;
in the formula (I), the compound is shown in the specification,wtfor the theoretical width of the steel plate after perspective transformation, wmask_pFor the calculated width of the mask region after perspective, wmask_aTo perspective the actual width of the front mask region, wplate_aIs the actual width of the steel plate;
if the width of the minimum coating rectangle is smaller than the theoretical width of the steel plate after perspective transformation, circularly reducing the threshold value according to the specified step length of 5-20 until the width of the minimum coating rectangle is larger than the theoretical width of the steel plate after perspective transformation; the process is coarse adjustment, the threshold value obtained at the end of the process is circularly increased by 1, until the width of the minimum coating rectangle is just less than or equal to the theoretical width of the steel plate after perspective transformation, and the process is fine adjustment;
repeating the process for each outline, namely the actual steel plate in sequence to obtain a self-adaptive threshold value and a steel plate rectangular outline meeting the actual steel plate width;
and calculating the actual position of the online steel plate according to the position of each rectangular outline and the actual size corresponding to the projected mask.
2. The method of claim 1, wherein the first adaptive threshold procedure in step (4) is as follows:
firstly, carrying out binarization by using an initial threshold value to obtain a rough area of a steel plate and other noise areas; traversing the outlines of all the areas and performing descending arrangement according to the areas of the areas, if the number of the found outlines is less than the number of the actual steel plates, circularly reducing the threshold value according to the set step length of 5-20 until the number of the detected outlines is more than or equal to the number of the actual steel plates; and marking the front actual steel plate number contours sorted according to the area after the number of the contours is larger than or equal to the actual steel plate number.
3. The method of claim 1 or 2, further comprising, after the step (5): and back projecting the rectangular outline of the steel plate to the original frame picture.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810879301.4A CN109190616B (en) | 2018-08-03 | 2018-08-03 | Hot-rolled steel plate online visual tracking method based on feature recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810879301.4A CN109190616B (en) | 2018-08-03 | 2018-08-03 | Hot-rolled steel plate online visual tracking method based on feature recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109190616A CN109190616A (en) | 2019-01-11 |
| CN109190616B true CN109190616B (en) | 2021-10-15 |
Family
ID=64920106
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810879301.4A Active CN109190616B (en) | 2018-08-03 | 2018-08-03 | Hot-rolled steel plate online visual tracking method based on feature recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109190616B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109816645B (en) * | 2019-01-18 | 2020-11-17 | 创新奇智(广州)科技有限公司 | Automatic detection method for steel coil loosening |
| CN110400321A (en) * | 2019-07-26 | 2019-11-01 | 广东工业大学 | The extracting method of leather material profile based on machine vision and actual size |
| CN111369554A (en) * | 2020-03-18 | 2020-07-03 | 山西安数智能科技有限公司 | Optimization and pretreatment method of belt damage sample in low-brightness multi-angle environment |
| CN113816098A (en) * | 2021-08-25 | 2021-12-21 | 南京钢铁股份有限公司 | Method for realizing full-automatic steel transferring of steel plate |
| CN115213241B (en) * | 2022-09-20 | 2023-02-03 | 山东钢铁股份有限公司 | Online identification method and system for thick hot steel plate |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101224472A (en) * | 2008-02-03 | 2008-07-23 | 东北大学 | Plate head bending shape detection device based on nearing fared image and method thereof |
| CN101574709A (en) * | 2009-06-12 | 2009-11-11 | 东北大学 | Automatic steel rotation method for medium plates |
| CN103267764A (en) * | 2013-05-30 | 2013-08-28 | 东北大学 | Hot-rolled steel plate surface defect image identification method based on neighborhood information estimation |
| CN103357672A (en) * | 2012-03-30 | 2013-10-23 | 鞍钢股份有限公司 | Online strip steel boundary detection method |
| CN103389733A (en) * | 2013-08-02 | 2013-11-13 | 重庆市科学技术研究院 | Vehicle line walking method and system based on machine vision |
| CN103861877A (en) * | 2014-03-27 | 2014-06-18 | 东北大学 | System and method for tracking and controlling position of steel plate of moderate-thickness plate thermal treatment furnace |
| CN104361353A (en) * | 2014-11-17 | 2015-02-18 | 山东大学 | Application of area-of-interest positioning method to instrument monitoring identification |
| CN104760812A (en) * | 2015-02-26 | 2015-07-08 | 三峡大学 | Monocular vision based real-time location system and method for products on conveying belt |
| CN105352437A (en) * | 2015-10-21 | 2016-02-24 | 广州视源电子科技股份有限公司 | Board card position detection method and device |
| CN107123188A (en) * | 2016-12-20 | 2017-09-01 | 北京联合众为科技发展有限公司 | Ticket of hindering based on template matching algorithm and edge feature is recognized and localization method |
| CN107589758A (en) * | 2017-08-30 | 2018-01-16 | 武汉大学 | A kind of intelligent field unmanned plane rescue method and system based on double source video analysis |
-
2018
- 2018-08-03 CN CN201810879301.4A patent/CN109190616B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101224472A (en) * | 2008-02-03 | 2008-07-23 | 东北大学 | Plate head bending shape detection device based on nearing fared image and method thereof |
| CN101574709A (en) * | 2009-06-12 | 2009-11-11 | 东北大学 | Automatic steel rotation method for medium plates |
| CN103357672A (en) * | 2012-03-30 | 2013-10-23 | 鞍钢股份有限公司 | Online strip steel boundary detection method |
| CN103267764A (en) * | 2013-05-30 | 2013-08-28 | 东北大学 | Hot-rolled steel plate surface defect image identification method based on neighborhood information estimation |
| CN103389733A (en) * | 2013-08-02 | 2013-11-13 | 重庆市科学技术研究院 | Vehicle line walking method and system based on machine vision |
| CN103861877A (en) * | 2014-03-27 | 2014-06-18 | 东北大学 | System and method for tracking and controlling position of steel plate of moderate-thickness plate thermal treatment furnace |
| CN104361353A (en) * | 2014-11-17 | 2015-02-18 | 山东大学 | Application of area-of-interest positioning method to instrument monitoring identification |
| CN104760812A (en) * | 2015-02-26 | 2015-07-08 | 三峡大学 | Monocular vision based real-time location system and method for products on conveying belt |
| CN105352437A (en) * | 2015-10-21 | 2016-02-24 | 广州视源电子科技股份有限公司 | Board card position detection method and device |
| CN107123188A (en) * | 2016-12-20 | 2017-09-01 | 北京联合众为科技发展有限公司 | Ticket of hindering based on template matching algorithm and edge feature is recognized and localization method |
| CN107589758A (en) * | 2017-08-30 | 2018-01-16 | 武汉大学 | A kind of intelligent field unmanned plane rescue method and system based on double source video analysis |
Non-Patent Citations (2)
| Title |
|---|
| "Visual Remote Monitoring and Control System for Rod Braking on Hot Rolling Mills";Oleg Starostenko 等;《Springer》;20171231;第297-307页 * |
| "基于机器视觉检测钢板板形的图像处理方法研究";许博文 等;《软件工程》;20180630;第21卷(第6期);第1-3页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109190616A (en) | 2019-01-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109190616B (en) | Hot-rolled steel plate online visual tracking method based on feature recognition | |
| CN109459750B (en) | Front multi-vehicle tracking method integrating millimeter wave radar and deep learning vision | |
| CN110097093B (en) | Method for accurately matching heterogeneous images | |
| CN104156726B (en) | A kind of workpiece identification method and device based on geometric characteristic | |
| CN101369346B (en) | Tracing method for video movement objective self-adapting window | |
| CN111563878B (en) | Space target positioning method | |
| EP0864134A2 (en) | Vector correlation system for automatically locating patterns in an image | |
| CN101826157B (en) | Ground static target real-time identifying and tracking method | |
| JPH10253322A (en) | Method and apparatus for designating position of object in space | |
| CN106709943B (en) | A kind of point cloud registration method based on optimal transmission | |
| CN106936964A (en) | A kind of mobile phone screen angular-point detection method based on Hough transformation template matches | |
| CN104391294B (en) | A kind of radar plot correlating method based on connected component and template matches | |
| CN114399675A (en) | Target detection method and device based on machine vision and laser radar fusion | |
| CN111127384A (en) | Strong reflection workpiece vision measurement method based on polarization imaging | |
| CN106600613A (en) | Embedded GPU-based improved LBP infrared target detection method | |
| CN106027911A (en) | In-orbit focusing method of earth observation satellite-borne visible light transmission camera | |
| CN109671084B (en) | Method for measuring shape of workpiece | |
| CN114387262A (en) | Nut positioning detection method, device and system based on machine vision | |
| CN114104655A (en) | Chain line following method and device | |
| CN113436239A (en) | Monocular image three-dimensional target detection method based on depth information estimation | |
| CN111203881B (en) | On-line feedback hand-eye calibration method based on weighted least square method | |
| KR101391667B1 (en) | A model learning and recognition method for object category recognition robust to scale changes | |
| Popov | Multicomponent Model of Objects Attributive Signatures on Color Pictures | |
| CN106408600A (en) | Image registration method applied to solar high-resolution image | |
| CN111626325A (en) | Image matching method based on features |
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 |