CN113139988B - Image processing method for efficiently and accurately estimating target scale change - Google Patents
Image processing method for efficiently and accurately estimating target scale change Download PDFInfo
- Publication number
- CN113139988B CN113139988B CN202110539550.0A CN202110539550A CN113139988B CN 113139988 B CN113139988 B CN 113139988B CN 202110539550 A CN202110539550 A CN 202110539550A CN 113139988 B CN113139988 B CN 113139988B
- Authority
- CN
- China
- Prior art keywords
- target
- scale
- template
- image
- estimated
- 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
- 238000003672 processing method Methods 0.000 title claims abstract description 14
- 239000013598 vector Substances 0.000 claims abstract description 43
- 230000009466 transformation Effects 0.000 claims abstract description 39
- 239000011159 matrix material Substances 0.000 claims abstract description 33
- 238000004364 calculation method Methods 0.000 claims abstract description 23
- 238000000034 method Methods 0.000 claims description 24
- 238000003491 array Methods 0.000 claims description 4
- 239000000654 additive Substances 0.000 claims description 3
- 230000000996 additive effect Effects 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000009499 grossing Methods 0.000 claims description 2
- 238000007781 pre-processing Methods 0.000 claims description 2
- 238000003384 imaging method Methods 0.000 description 2
- 241000566145 Otus Species 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/181—Segmentation; Edge detection involving edge growing; involving edge linking
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- 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
- G06T2207/20024—Filtering details
- G06T2207/20032—Median filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an image processing method for efficiently and accurately estimating target scale change, which intercepts a target circumscribed rectangular area for calculation so as to reduce time consumption and improve rough estimation precision. And carrying out logarithmic transformation on the target coordinates to convert the multiplication calculation of scale change into addition calculation. And storing the target contour of the original image in a sparse matrix form after coordinate transformation, and taking the target contour as a template. And adding the template and the scale vector and reducing the template and the scale vector into a binary image, matching the binary image with the image to be estimated, and searching for an optimal matching point to obtain a final estimation result.
Description
Technical Field
The invention relates to the technical field of scale change estimation of targets at different distances in target tracking, in particular to an image processing method for efficiently and accurately estimating scale change of a target.
Background
When a target is tracked, the scale of a moving target changes along with the change of the distance in the imaging of a detector, a tracking wave gate needs to be changed in time to adapt to the scale change of the target, and high-precision scale estimation is needed to adaptively adjust the parameters of a tracking algorithm, so that the moving target is stably tracked.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the method overcomes the defects of the prior art and provides an image processing method for estimating the target scale change with high efficiency and high accuracy.
The technical scheme adopted by the invention for solving the technical problems is as follows:
an image processing method for efficiently and accurately estimating target scale change comprises the following steps: after gradient of the image with the target is solved and binarization processing is carried out, intercepting a rectangular region externally connected with the target for calculation; respectively carrying out logarithmic transformation on the horizontal and vertical coordinates of the template and the circumscribed rectangle of the target to be estimated by taking the upper left corner of the circumscribed rectangle as an origin, and converting the multiplicative calculation of the scale change of the target to be estimated relative to the target template into the additive calculation of the scale of the target template and the transformation offset; carrying out logarithmic transformation on the target contour coordinates of the original image, storing the target contour coordinates in a sparse matrix form, and taking the sparse matrix form as a template; and adding the template and elements in the scale vector and reducing the template and the elements in the scale vector into a binary image, matching the binary image with the target feature to be estimated, traversing the scale vector and searching for optimal scale change data.
The method is characterized in that logarithmic transformation is used for converting multiplication calculation of scale change into addition calculation, only an external rectangular area is processed, the upper left corner of the external rectangle is used as an origin, logarithmic transformation is respectively carried out on the transverse coordinates and the longitudinal coordinates of the external rectangle of the template and the target area to be estimated, and then the scale after transformation of the target to be estimated is converted into template transformation scale and offset.
The sparse matrix stores target contour coordinates subjected to logarithmic transformation in a sparse matrix form; and (4) carrying out logarithmic transformation on the left boundary, the right boundary and the corresponding longitudinal axis coordinate of the target contour, forming a cluster by the three data, and combining all clusters to form a sparse matrix.
And discretizing the data by a certain step length in the possible range of the scale to obtain a group of possible scale estimation arrays, and carrying out logarithmic transformation on the scale estimation arrays to obtain the scale vectors.
And the target feature to be estimated is obtained by solving the gradient of the image to be estimated, selecting a proper threshold value for binarization processing, and intercepting a target circumscribed rectangular area as the target feature to be estimated.
The method comprises the steps of adding the template and elements in the scale vector, reducing the added coordinates into a binary image through exponential operation, matching the binary image with target features to be estimated, carrying out AND operation on the reduced binary image and the target features to be estimated, counting the number of points with pixels of 1 after the AND operation, and taking the points as matching values under corresponding scale elements.
Traversing the scale vector, and searching optimal scale change data; the operation in claim 6 is performed on the elements of each scale vector, after traversing the scale vectors is finished, the point with the maximum matching value is searched as the optimal matching point, and the scale elements are subjected to exponential transformation to obtain the final scale change data.
A high-efficient high-accuracy image processing method of estimating the change of target dimension, the moving target changes with the distance, the size change of imaging takes place in the detector, adopt this method, estimate out the scale change multiple k compared with original target rapidly, this method process is as follows, the pixel coordinate of the picture is (x, y), after amplifying the picture k times, the corresponding pixel coordinate changes to (k x, k y), in order to reduce the calculated amount, only use the coordinate calculation of the target outline area; preprocessing an original image and a changed image, extracting an external rectangle of a target outline, comparing the sizes of the external rectangles of the original target and the changed target to obtain a rough scale change estimation, carrying out logarithmic transformation on coordinates of the external rectangle, converting the scale change from multiplication operation to addition operation, searching an optimal translation point by using a translation searching mode, and carrying out corresponding index transformation to obtain final scale change data.
An image processing method for efficiently and accurately estimating target scale change comprises the following specific implementation steps:
first, the original target is preprocessed. And smoothing the original image with the target to reduce noise influence, calculating the gradient of the smoothed image through a gradient operator, and selecting a proper threshold value to binarize the gradient image.
And secondly, constructing a target template.
The target template is stored in the form of a sparse matrix. The present invention defines the stored sparse matrix as follows: the binarized image only contains target elements with pixel values of 1 and background elements with pixel values of 0, so that the pixel values of the target are ignored, the sparse matrix only stores horizontal and vertical direction coordinates of the target elements, and the target pixels are all continuous and clustered, so that the sparse matrix is stored in a clustered mode, and each cluster comprises three elements: and carrying out logarithmic transformation on the three elements according to the initial coordinate and the final coordinate in the horizontal direction and the corresponding coordinate in the vertical direction. The set of all blobs constitutes a sparse matrix.
And intercepting a target area in the binary image, calculating a circumscribed rectangle of the target, taking the upper left corner of the circumscribed rectangle as an origin, and only carrying out logarithmic transformation on the coordinates of the circumscribed rectangle area again to obtain a target contour sparse matrix to construct a target template for matching.
Thirdly, processing the target area to be estimated
And storing the circumscribed rectangular area of the target to be estimated as a new image, and obtaining an area binary image by adopting a gradient calculation and threshold segmentation method which is the same as the template processing method for the image.
And fourthly, roughly estimating the scale.
And dividing the length and the width of the circumscribed rectangle of the target to be estimated by the length and the width of the target template respectively to obtain the size of the rough estimation scale.
Fifthly, scale estimation and fine calculation
And searching for the optimal scale estimation by adopting a traversal search method. Selecting possible values in the neighborhood of the rough estimation scale, taking the rough estimation scale as a center, possible scaling scales as radii and preset scale precision as step length, constructing an enumeration search vector, and carrying out logarithmic transformation on the vector.
Circularly taking the search elements from the search vector, and performing the following operation on each search element:
and adding target template elements and search elements one by one to obtain a direction superposition amount corresponding to a search scale, performing corresponding exponential transformation on the direction superposition amount to reduce the direction superposition amount into a binary image, performing and operation on corresponding pixel points of the area binary image to obtain a matched binary image, and counting the number of effective points of the matched binary image after all elements of the current search vector are calculated to serve as a matching coefficient of the current search element.
And after traversing the search vector, counting the matching coefficient corresponding to each element in the vector, wherein the element with the highest matching coefficient is the optimal matching point, performing index transformation on the element, and obtaining a result which is the final scale estimation result.
Compared with the prior art, the invention has the advantages that:
(1) The invention intercepts the external rectangle of the target for calculation, reduces the calculation range, gives the rough estimation amount of the scale change by comparing the sizes of the external rectangles, and lays a cushion for reducing the calculation amount of the traversal estimation;
(2) According to the method, the target edge information is stored in a sparse matrix form, only the target effective data is processed, and compared with the traditional method of extracting features in a full image, the calculated amount generated by a non-target area is eliminated;
(3) The method carries out logarithmic transformation on effective coordinate data screened by the circumscribed rectangle and the sparse matrix, matches the data to be estimated with template data by adding discretized offset, and has the matching mode of AND operation of corresponding pixels.
(4) The final estimation precision of the method is related to the precision of the search vector, and the precision of the search vector can be adjusted according to actual requirements to achieve the expected estimation precision. Whereas the estimation results of the prior art can only output one precision.
Drawings
FIG. 1 is a flowchart of an image processing method for efficiently and accurately estimating a target scale change according to the present invention;
FIG. 2 is a binary image I of the present invention obtained by adding vector elements to a template r And (5) making a matching flow chart.
Detailed Description
The invention is further described with reference to the following figures and detailed description.
As shown in fig. 1, the image processing method for efficiently and accurately estimating the target scale change of the present invention specifically includes the following steps:
in the first step, 3*3 are usedFiltering the original image by a value filter or a median filter, if necessary, increasing the kernel size of the filter to reduce the influence of noise on the processing result, calculating the gradient of the smoothed image by using a gradient operator, calculating a segmentation threshold value T for the gradient map by using an Otus method, and binarizing the gradient map by using the T to obtain a binarized image I BW 。
And secondly, constructing a target template.
The target template is stored in the form of a sparse matrix. The definition of the sparse matrix in the invention is as follows: the target appears in each row of the image in a segmented manner, the left boundary pixel coordinate of each segment of data is xl, the right boundary pixel coordinate of each segment of data is xr, the row coordinate of the target is y, each segment of data forms a group of clusters to form a group of elements of a sparse matrix, and all data segments of the target in the image form a complete sparse matrix. The Sparse matrix Sparse [ i ] [ j ] is represented by a two-dimensional matrix: the value of i is 0-2, which respectively represents xl, xr and y, j is the jth group. Sparse [0] [ j ], sparse [1] [ j ], sparse [2] [ j ] denote the coordinate positions of xl, xr, y of the jth blob, respectively.
Intercepting circumscribed rectangles of the target in the binary image, wherein the length and the width of each circumscribed rectangle are L, W, the left upper corner of each circumscribed rectangle is used as an origin, only carrying out logarithmic transformation on the coordinates of the region of the circumscribed rectangle again, carrying out logarithmic transformation on the coordinates of each cluster to obtain log2 (xl), log2 (xr) and log2 (y), and constructing a target template Sparse _ base for matching by the transformed Sparse matrix.
Thirdly, processing the target area to be estimated
Saving the circumscribed rectangular area of the target to be estimated as a new image I n To 1, pair n Obtaining a regional binary image I by adopting a first-step processing method r 。
Fourth, coarse scale estimation
The length and width of a circumscribed rectangle of the target to be estimated are L and W respectively, and the rough estimation scale size K = L/L = W/W.
Fifthly, scale estimation and fine calculation
Constructing a fine scale enumeration vector V = { K-2,K-2+ step, K-2+2 step, … …, K +2} in a coarse estimation scale neighborhood, wherein step isp is the precision of the scale to be estimated, and the value can be 0.2, 0.1, 0.05, 0.01 and the like. Carrying out logarithmic transformation on the scale enumeration vector to obtain a search vector V log ＝{log2(K-2),log2(K-2+step),log2(K-2+2*step),……，log(K+2)}。
Cyclically taking search elements from the search vector, and performing the following operations on each search element:
target template elements are taken one by one from a group of Sparse matrix Sparse _ base [ n ] = (log (xl (n)), log (xr (n)), log (y (n))), and added with search elements to obtain a direction superposition amount corresponding to the search scale, and corresponding exponential transformation is carried out to obtain [ xs, yl ] = [2^ (log (xl (n)) + LK) + m,2^ (log (y (n)) + LK) ],
where m is an integer starting from 0, accumulated until xs = =2^ (log (xr (n)) + LK);
LK＝V log (p) = log (K-2 + p step), i.e. one element in the search vector.
Restoring to the corresponding position in the binary image, corresponding pixel position 1,I S (xs, yl) =1, and is associated with the region binary image I r And operation is carried out on corresponding pixel points to obtain a matched binary image I C (xs,yl)＝I S (xs,yl)&I r (xs, yl), after all elements of the current search vector are calculated, counting the number of effective points of the matched binary image:
as the matching coefficient for the current search element.
After traversing the search vector, counting the matching coefficient corresponding to each element in the vector, and counting the element V with the maximum matching coefficient N log (p) is the optimal matching point, the element is subjected to exponential transformation, and the obtained result K = K-2+ p step is the final scale estimation result.
Claims (7)
1. An image processing method for efficiently and accurately estimating target scale change is characterized in that: the method comprises the following steps: after gradient of the image with the target is solved and binarization processing is carried out, intercepting a rectangular region externally connected with the target for calculation; taking the upper left corner of the circumscribed rectangle as an original point, respectively carrying out logarithmic transformation on the transverse coordinates and the longitudinal coordinates of the circumscribed rectangle of the template and the target to be estimated, and converting the multiplication calculation of the scale change of the target to be estimated relative to the target template into addition calculation; carrying out logarithmic transformation on the target contour coordinates of the original image, storing the target contour coordinates in a sparse matrix form, and taking the sparse matrix form as a template; adding the template and elements in the scale vector and reducing the template and the elements in the scale vector into a binary image, matching the binary image with the target feature to be estimated, traversing the scale vector and searching for optimal scale change data;
the sparse matrix stores target contour coordinates subjected to logarithmic transformation in a sparse matrix form; and (4) carrying out logarithmic transformation on the left boundary, the right boundary and the corresponding longitudinal axis coordinate of the target contour, forming a cluster by the three data, and combining all clusters to form a sparse matrix.
2. The method as claimed in claim 1, wherein the multiplicative calculation for the target to be estimated with respect to the target template scale change is converted into additive calculation, only the circumscribed rectangle region is processed, the upper left corner of the circumscribed rectangle is used as the origin, the horizontal and vertical coordinates of the circumscribed rectangles of the template and the target region to be estimated are respectively subjected to logarithmic transformation, and the transformed scale of the target to be estimated is converted into additive calculation of the template transformation scale plus the offset.
3. The method as claimed in claim 1, wherein the scale vector is obtained by discretizing data in a scale possible range by a certain step size to obtain a set of possible scale estimation arrays, and performing logarithmic transformation on the scale estimation arrays to obtain the scale vector.
4. The method for processing the image with the high efficiency and the high accuracy for estimating the target scale change according to claim 1, wherein the target feature to be estimated is characterized in that a gradient of the image to be estimated is obtained, a proper threshold value is selected for binarization processing, and a target circumscribed rectangular region is intercepted as the target feature to be estimated.
5. The method for processing the image with high efficiency and high accuracy for estimating the scale change of the target according to claim 1, wherein the template and the elements in the scale vector are added and restored to the binary image, the method comprises the steps of adding the template and the elements in the scale vector, restoring the added coordinates to the binary image through exponential operation, and matching the binary image with the target feature to be estimated, and the method comprises the steps of performing and operation on the restored binary image and the target feature to be estimated, and counting the number of points of which the pixels are 1 after the sum operation, wherein the points are used as the matching value under the corresponding scale element.
6. The method for efficiently and accurately estimating the scale change of the target according to claim 5, wherein the scale vector is traversed to find optimal scale change data; the operation in claim 5 is performed on the elements of each scale vector, after traversing the scale vectors is finished, the point with the maximum matching value is searched as the optimal matching point, and the scale elements are subjected to exponential transformation to obtain the final scale change data.
7. An image processing method for efficiently and accurately estimating target scale change is characterized in that: the method comprises the following concrete steps:
firstly, preprocessing an original target: smoothing the original image with the target to reduce noise influence, calculating the gradient of the smoothed image through a gradient operator, and selecting a proper threshold value to binarize the gradient image;
second, construct the target template
Storing the target template in a sparse matrix form, wherein the stored sparse matrix is defined as follows: the binarized image only contains target elements with the pixel value of 1 and background elements with the pixel value of 0, so that the pixel value of the target is ignored, the sparse matrix only stores the horizontal and vertical direction coordinates of the target elements, and the target pixels are all continuous and clustered, so that the sparse matrix is stored in a clustered mode, wherein each cluster comprises three elements: carrying out logarithmic transformation on the three elements according to the initial coordinate and the final coordinate in the horizontal direction and the coordinate corresponding to the vertical direction, and forming a sparse matrix by the set of all clusters;
intercepting a target area in the binary image, calculating a circumscribed rectangle of the target, taking the upper left corner of the circumscribed rectangle as an origin, and only carrying out logarithmic transformation on the coordinates of the circumscribed rectangle area again to obtain a target contour sparse matrix to construct a target template for matching;
thirdly, processing the target area to be estimated
Storing the circumscribed rectangular area of the target to be estimated as a new image, and obtaining an area binary image by adopting a gradient calculation and threshold segmentation method which is the same as template processing on the image;
fourth, coarse scale estimation
Dividing the length and the width of the circumscribed rectangle of the target to be estimated by the length and the width of the target template respectively to obtain the size of a rough estimation scale;
fifthly, scale estimation and fine calculation
Searching for optimal scale estimation by adopting a traversal search method, selecting possible numerical values in a coarse estimation scale neighborhood, constructing an enumeration search vector by taking the coarse estimation scale as a center, possible scaling scales as radii and preset scale precision as a step length, and carrying out logarithmic transformation on the vector;
cyclically taking search elements from the search vector, and performing the following operations on each search element:
adding target template elements and search elements one by one to obtain a direction superposition amount corresponding to a search scale, performing corresponding exponential transformation on the direction superposition amount to reduce the direction superposition amount into a binary image, performing and operation on pixel points corresponding to the regional binary image to obtain a matched binary image, and counting the number of effective points of the matched binary image after all elements of the current search vector are calculated to serve as a matching coefficient of the current search element;
and after traversing the search vector, counting the matching coefficient corresponding to each element in the vector, wherein the element with the highest matching coefficient is the optimal matching point, performing index transformation on the element, and obtaining a result which is the final scale estimation result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110539550.0A CN113139988B (en) | 2021-05-17 | 2021-05-17 | Image processing method for efficiently and accurately estimating target scale change |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110539550.0A CN113139988B (en) | 2021-05-17 | 2021-05-17 | Image processing method for efficiently and accurately estimating target scale change |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113139988A CN113139988A (en) | 2021-07-20 |
| CN113139988B true CN113139988B (en) | 2023-02-14 |
Family
ID=76817219
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110539550.0A Active CN113139988B (en) | 2021-05-17 | 2021-05-17 | Image processing method for efficiently and accurately estimating target scale change |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113139988B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7391423B1 (en) * | 2004-10-06 | 2008-06-24 | Adobe Systems Incorporated | Thumbnail scaling based on display pane size |
| CN103632132A (en) * | 2012-12-11 | 2014-03-12 | 广西工学院 | Face detection and recognition method based on skin color segmentation and template matching |
| CN104463866A (en) * | 2014-12-04 | 2015-03-25 | 无锡日联科技有限公司 | Local shape matching method based on outline random sampling |
| CN108052942A (en) * | 2017-12-28 | 2018-05-18 | 南京理工大学 | A kind of visual pattern recognition methods of aircraft flight attitude |
| CN108256394A (en) * | 2016-12-28 | 2018-07-06 | 中林信达（北京）科技信息有限责任公司 | A kind of method for tracking target based on profile gradients |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10019631B2 (en) * | 2015-11-05 | 2018-07-10 | Qualcomm Incorporated | Adapting to appearance variations when tracking a target object in video sequence |
| CN109345583B (en) * | 2018-10-11 | 2020-10-09 | 西安电子科技大学 | SAR target image geometric dimension estimation method based on OMP |
-
2021
- 2021-05-17 CN CN202110539550.0A patent/CN113139988B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7391423B1 (en) * | 2004-10-06 | 2008-06-24 | Adobe Systems Incorporated | Thumbnail scaling based on display pane size |
| CN103632132A (en) * | 2012-12-11 | 2014-03-12 | 广西工学院 | Face detection and recognition method based on skin color segmentation and template matching |
| CN104463866A (en) * | 2014-12-04 | 2015-03-25 | 无锡日联科技有限公司 | Local shape matching method based on outline random sampling |
| CN108256394A (en) * | 2016-12-28 | 2018-07-06 | 中林信达（北京）科技信息有限责任公司 | A kind of method for tracking target based on profile gradients |
| CN108052942A (en) * | 2017-12-28 | 2018-05-18 | 南京理工大学 | A kind of visual pattern recognition methods of aircraft flight attitude |
Non-Patent Citations (2)
| Title |
|---|
| 光电成像目标识别与检测关键技术研究;钟剑丹;《中国优秀博士学位论文全文数据库 信息科技辑》;20181015;全文 * |
| 基于Harris角点与广义霍夫变换的扩展目标稳定跟踪算法;张礼，魏宇星等;《国外电子测量技术》;20190915;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113139988A (en) | 2021-07-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110866924B (en) | Line structured light center line extraction method and storage medium | |
| CN110335234B (en) | Three-dimensional change detection method based on antique LiDAR point cloud | |
| WO2019042232A1 (en) | Fast and robust multimodal remote sensing image matching method and system | |
| CN111462120B (en) | Defect detection method, device, medium and equipment based on semantic segmentation model | |
| CN107240130B (en) | Remote sensing image registration method, device and system | |
| CN103337077A (en) | Registration method for visible light and infrared images based on multi-scale segmentation and SIFT (Scale Invariant Feature Transform) | |
| CN111523547B (en) | 3D semantic segmentation method and terminal | |
| CN113436207A (en) | Method for quickly and accurately extracting line structure light stripe center of regular surface | |
| CN108764343B (en) | Method for positioning tracking target frame in tracking algorithm | |
| Chen et al. | A novel Fourier descriptor based image alignment algorithm for automatic optical inspection | |
| CN113139988B (en) | Image processing method for efficiently and accurately estimating target scale change | |
| Wu et al. | Research on crack detection algorithm of asphalt pavement | |
| CN101430789A (en) | Image edge detection method based on Fast Slant Stack transformation | |
| You et al. | A guided image matching approach using Hausdorff distance with interesting points detection | |
| WO2019041447A1 (en) | 3d video frame feature point extraction method and system | |
| CN117078726A (en) | Different spectrum image registration method based on edge extraction | |
| CN109829502B (en) | Image pair efficient dense matching method facing repeated textures and non-rigid deformation | |
| Giraud et al. | Texture superpixel clustering from patch-based nearest neighbor matching | |
| Lindblad et al. | Exact linear time euclidean distance transforms of grid line sampled shapes | |
| CN115620263A (en) | Intelligent vehicle obstacle detection method based on image fusion of camera and laser radar | |
| CN112883796B (en) | SAR image multi-type building height estimation method based on overlapping and masking information | |
| CN112651936A (en) | Steel plate surface defect image segmentation method and system based on image local entropy | |
| Sha et al. | A fast matching algorithm based on K-degree template | |
| Murashov | Method for combining image segmentation maps on the basis of information redundancy and variation of information minimization | |
| CN110603535A (en) | Iterative multi-directional image search supporting large template matching |
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 |