CN113470001A - Target searching method for infrared image - Google Patents
Target searching method for infrared image Download PDFInfo
- Publication number
- CN113470001A CN113470001A CN202110830767.7A CN202110830767A CN113470001A CN 113470001 A CN113470001 A CN 113470001A CN 202110830767 A CN202110830767 A CN 202110830767A CN 113470001 A CN113470001 A CN 113470001A
- Authority
- CN
- China
- Prior art keywords
- target
- image
- infrared image
- error
- graph
- 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 34
- 230000008569 process Effects 0.000 claims description 13
- 238000009825 accumulation Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000013461 design Methods 0.000 claims description 5
- 238000007781 pre-processing Methods 0.000 claims description 4
- 230000002708 enhancing effect Effects 0.000 claims description 3
- 230000003252 repetitive effect Effects 0.000 claims description 2
- 239000000126 substance Substances 0.000 claims 1
- 238000012545 processing Methods 0.000 abstract description 9
- 230000008859 change Effects 0.000 abstract description 4
- 230000000694 effects Effects 0.000 description 11
- 238000010586 diagram Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 2
- 238000012850 discrimination method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000006467 substitution reaction Methods 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
- G06T7/0002—Inspection of images, e.g. flaw detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/90—Dynamic range modification of images or parts thereof
-
- 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
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10048—Infrared 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
- G06T2207/20016—Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
-
- 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]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Quality & Reliability (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a target searching method for an infrared image, which belongs to the field of image processing and comprises the following steps: acquiring an infrared image, and performing image enhancement pretreatment on the infrared image to obtain a first layer of an image characteristic pyramid; manually marking a target position on the infrared image, namely giving a target frame of a first frame; extracting images of different levels in the target frame by using a characteristic pyramid algorithm on the infrared image: and (4) searching the target by utilizing an SSDA template matching algorithm in each layer, selecting the maximum possible position of the target, and simultaneously recording the confidence degree and the target moving position information to obtain the searched target. The invention utilizes the image characteristic pyramid algorithm to accelerate the searching speed, improve the frame rate, ensure that the effective characteristics of the target can be still stored when the target is subjected to scale change, and can complete the real-time accurate searching of the infrared image of the moving background.
Description
Technical Field
The invention belongs to the field of image processing, and particularly relates to a target searching method for an infrared image.
Background
With the development of intelligence, images become an important means for acquiring information. In image acquisition, image processing is a very important link, and image processing technology has been widely used in various fields such as industrial safety medical management. In image processing, searching for a specific object from among moving objects is a very important step in image processing.
When the infrared image of the moving background is searched in real time, the range needs to be locked as soon as possible to search and acquire the target, and the search is based on the processing of the infrared image. In the above situation, the image data amount is large, and it is difficult for the conventional image processing method to accurately search for the target.
Therefore, the invention provides a target searching method for an infrared image.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention provides a target searching method for an infrared image.
In order to achieve the above purpose, the invention provides the following technical scheme:
a target searching method for an infrared image comprises the following steps:
acquiring an infrared image, and performing image enhancement pretreatment on the infrared image to obtain a first layer of an image characteristic pyramid;
manually marking a target position on the infrared image, namely giving a target frame of a first frame;
extracting images of different levels in a target frame on the infrared image by using a characteristic pyramid algorithm:
aiming at the images of different levels obtained through the characteristic pyramid, target searching is carried out on each level by utilizing an improved SSDA template matching algorithm, the maximum possible position of a target is selected, and meanwhile, confidence and target moving position information are recorded, so that the searched target is obtained.
Preferably, the infrared image is a single-channel 8-bit image, and the pixel value of the image is xinThe image enhancement preprocessing comprises: traversing the image, enhancing the gray value of each pixel, and outputting the pixel valueIs xout:
Wherein k is the enhancement ratio.
Preferably, extracting images of different levels on the infrared image by using a feature pyramid algorithm means that convolution checks of the target by using convolution checks of 2x2, 3x3, 5x5 and 7x7 respectively to obtain fuzzy images of different degrees respectively; different target features are reflected on different feature images, and image information of different dimensions is detected from images of different scales.
Preferably, assuming that S (x, y) is a search graph of MxN and T (x, y) is a template graph of MxN, the target search process specifically includes:
error definition:
wherein S isi,jIs a subgraph of the search graph, T is a template graph,is the average value of the template graph, the initial position of the upper left corner of the subgraph is (i, j), and then the average subgraphIs as follows;
setting an initial threshold Th0, and when the threshold is matched with a random point in the SSDA algorithm, directly discarding the point if the point is deemed not to be a target by an error accumulation design threshold; the threshold is set to an empirical value;
randomly selecting non-repetitive pixel points in the area to be searched as the center of the tracking frame; calculating the current error, accumulating the error, and recording the current accumulation frequency H when the error accumulation exceeds a threshold Th 0; then traversing all the subgraphs;
in the traversing process, if the error value is greater than the threshold Th0 under the times less than H, the operation of selecting random points to calculate the error is not continued, and the next sub-graph is directly switched to;
in the traversing process, if a subgraph exists, after H times of calculation, the accumulated error is Th 1; if Th1< Th0, update Th0 to Th 1;
recording the matching times H and the accumulated error sum in all the sub-graph matching in the traversal process, and calculating the average error; and after all the subgraphs are traversed, outputting the average error rate and the minimum center coordinate of the subgraph to obtain a search target.
The target searching method for the infrared image provided by the invention has the following beneficial effects:
the invention utilizes the image characteristic pyramid algorithm to carry out different convolution fuzzy operations on the target, obtains the characteristic graphs of the image with different scales, carries out the matching algorithm on the characteristic graphs, can accelerate the searching speed and improve the frame rate, simultaneously can respectively calculate the values of the matching functions of each layer of the characteristic pyramid when the scale of the target changes, selects the optimal matching characteristic graph and carries out template switching, ensures that the effective characteristics of the target can still be stored when the scale of the target changes, and can complete the real-time accurate searching of the infrared image of the moving background.
Drawings
In order to more clearly illustrate the embodiments of the present invention and the design thereof, the drawings required for the embodiments will be briefly described below. The drawings in the following description are only some embodiments of the invention and it will be clear to a person skilled in the art that other drawings can be derived from them without inventive effort.
Fig. 1 is a flowchart of a target search method for infrared images according to embodiment 1 of the present invention;
FIG. 2 is an infrared image contrast enhancement map;
FIG. 3 is a diagram illustrating an image structure for extracting different levels using a feature pyramid;
FIG. 4 is a diagram of the effect of processing an infrared image directly using a Robert convolution kernel;
FIG. 5 is a diagram illustrating the effect of processing an infrared image after superimposing the results of two convolution kernels;
FIG. 6 is a diagram of the effect of the superimposed image after the small noise is removed;
FIG. 7 is a diagram of the effect of convolving an enhanced image by 3x 3;
FIG. 8 is a diagram of the effect of using the modified Laplacian on the enhanced image;
FIG. 9 is a diagram illustrating the effect of expanding the area of the field of view while reducing the weight parameters by using a plurality of convolution kernels in a superposition manner;
FIG. 10 is a graph of the effect of 5x5 and 7x7 convolutions performed on layer 3 and layer 4, respectively;
FIG. 11 is a diagram showing the change in the respective images when the target is subjected to a change in scale;
FIG. 12 is a search diagram;
fig. 13 is a flowchart of random matching.
Detailed Description
In order that those skilled in the art will better understand the technical solutions of the present invention and can practice the same, the present invention will be described in detail with reference to the accompanying drawings and specific examples. The following examples are only for illustrating the technical solutions of the present invention more clearly, and the protection scope of the present invention is not limited thereby.
Example 1
The invention provides a target searching method for an infrared image, which specifically comprises the following steps as shown in figure 1:
and S1, acquiring the infrared image, and performing image enhancement pretreatment on the infrared image to obtain the first layer of the image characteristic pyramid.
The infrared image reflects the temperature distribution of the image, and generally, the tracking target is an object with obvious difference between the temperature and the surrounding environment. In order to make the target more obvious, aiming at the infrared image, image enhancement is carried out, so that the realization of a subsequent tracking algorithm is facilitated.
In this embodiment, the infrared image is a single-channel 8-bit image, and the pixel value of the image is xinThe enhanced preprocessing of the infrared image comprises the following steps: traversing the image, enhancing the gray value of each pixel, and outputting the pixel value of xout:
Where k is the enhancement ratio, empirically chosen to be 1.3.
The enhanced image effect is shown in fig. 2, and it can be seen that the object is distinguished from the background more obviously.
S2, the target position is manually marked on the infrared image, i.e. the target box (ground route box) of the given first frame.
And S3, extracting images of different enhancement layers on the infrared image by using a characteristic pyramid algorithm.
Extracting images of different levels on the infrared image by using a feature pyramid algorithm means that convolution cores of 2x2, 3x3, 5x5 and 7x7 are respectively used for carrying out convolution on a target to respectively obtain fuzzy images of different degrees. Different target features can be reflected on different feature images, image information of different dimensions can be detected from images of different dimensions, and an image structure is shown in fig. 3.
The resolution of the infrared camera used in the present invention is 640x 512.
1. Layer 0-enhanced image
The image is enhanced as layer 0. Its size remains the same as the original image, 640x 512.
2. Layer 1-2 x2 convolution image
The 2x2 convolution image uses a modified version of the Robert convolution kernel with a step size of 1, padding is used, so its output size image is consistent with the original image, 640x 512. The Robert convolution kernel, as a first order differential, has small calculation amount and is sensitive to details.
The Robert convolution kernel contains the following two convolution kernels:
because the infrared image is used in the invention, the whole brightness of the obtained image is reduced after the infrared image is directly used, and the effect is not ideal, as shown in fig. 4.
Therefore, the structure is changed into the following structure,
and the results of the two convolution kernels are superimposed, the effect of which is shown in fig. 5:
the final superposed image is used as the opening operation of the primary image, small noise points are removed, the effect is shown in fig. 6, the image quality is clearer, and the main outline characteristics of the target are obvious.
3. Layer 2-3 x3 convolution image
The enhanced image is convolved 3x3, where the convolution kernel uses the modified laplacian, the original laplacian being:
the overall brightness of the result of such convolution is low, as shown in fig. 7:
hence the modified laplacian is used:
the improved image brightness was significantly enhanced, and the result is shown in fig. 8.
It can be seen that the image quality detail information after the 3x3 convolution is richer.
Since the step distance adopted by the 3 × 3 convolution is 2, the formula is calculated according to the image size after convolution:
where the size of the input image is H x W, the convolution kernel size is FH x FW, the stride is S, and the padding (padding) is P.
Thus, after convolution by 3x3 and cropping out the surrounding pixels, the resulting image size is 320 x 256.
When the size of the image is smaller and the template matching is carried out in the later period, the target searching speed is accelerated, and the speed is accelerated.
4. Layer 3-5 x5 convolution image
5. Layer 4-7 x7 convolution image
The 3 rd layer and the 4 th layer are respectively convoluted by 5x5 and 7x7, but the conventional 5x5 and 7x7 convolution methods pose certain difficulty for padding calculation and step size programming, and the number of parameters is large, and if a reasonable convolution kernel is designed, a large amount of time is wasted, so that the same receptive field is obtained by using multiple 3x3 convolution kernels by taking the advantage of the convolution kernel accumulation method in deep learning, and a schematic diagram of the method is shown in fig. 9.
The image subjected to convolution of 5x5 is cut to have a size of 160x128, and the image subjected to convolution of 7x7 is cut to have a size of 80x 64. The convolved image is shown in fig. 10 (zoomed in to the same view for easy viewing, not actual size).
By the method, when the target changes from far to near, the rough position can be judged without researching detailed information, and the matching method is comprehensively designed by comprehensively utilizing the characteristics and advantages of each image through a later error function discrimination method.
The situation on each image when the object undergoes a change in scale is shown in fig. 11.
And S4, performing an improved SSDA template matching algorithm on the images of different layers obtained by the characteristic pyramid in the S3, selecting the maximum possible position of the target, and recording the confidence degree and the target moving position information.
Assuming that S (x, y) is a search map of M × N, T (x, y) is a template map of M × N, Si,jA sub-graph (the initial position at the upper left corner is (i, j)) in the search graph completes the search by sliding on the graph to be searched, as shown in fig. 12, the target search process specifically includes:
s4.1, error definition:
wherein S isi,jIs a subgraph of the search graph, T is a template graph,is the average value of the template graph, the initial position of the upper left corner of the subgraph is (i, j), and then the average subgraphIs as follows;
s4.2, setting an initial threshold Th0, and when the threshold is matched with a random point in the SSDA algorithm, directly discarding the point if the point is deemed not to be a target by the error accumulation design threshold. The threshold is set to an empirical value, typically 30-40.
S4.3, random matching method
And randomly selecting non-repeated pixel points in the area to be searched as the center of the tracking frame. The current error is calculated and accumulated, and when the error accumulation exceeds a threshold Th0, the current accumulation number H is recorded. All subgraphs are then traversed.
In the traversal process, if the error value is greater than the threshold Th0 for times less than H, the operation of selecting a random point to calculate the error is not continued, and the switching to the next sub-graph is directly performed.
In the traversal process, if there is one sub-graph, when H times are calculated, the accumulated error is Th 1. If Th1< Th0, Th0 is updated to Th 1.
In order to ensure the false detection caused by the accidental situation in the random calculation process, a lower limit should be set for the random number H, which is generally not less than 40% of the sub-pixel number.
And recording the matching times H and the accumulated error sum in all the sub-graph matching in the traversal process, and calculating the average error sum Ea. After all the subgraphs are traversed, the average error rate and the minimum center coordinate of the subgraph are output, the flow chart of random matching is shown in fig. 13, the lines corresponding to a, b, c and d in the graph (a) and the lines corresponding to a, b, c and d in the graph (b) are one line respectively.
The method aims at the heat distribution rule of the infrared image, carries out nonlinear image enhancement preprocessing, strengthens the characteristics of a target, utilizes an image characteristic pyramid (FPN) algorithm, carries out an improved SSDA template matching algorithm on a plurality of layers, designs an error function discrimination method and finds a proper template updating time. When the error discrimination function judges that the target is lost, predicting the reappearance approximate range of the target according to the central coordinate position of the tracking frame of the frames of pictures before the target is lost, and adaptively changing the search domain.
The above-mentioned embodiments are only preferred embodiments of the present invention, and the scope of the present invention is not limited thereto, and any simple modifications or equivalent substitutions of the technical solutions that can be obviously obtained by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.
Claims (4)
1. A target searching method for an infrared image is characterized by comprising the following steps:
acquiring an infrared image, and performing image enhancement pretreatment on the infrared image to obtain a first layer of an image characteristic pyramid;
manually marking a target position on the infrared image, namely giving a target frame of a first frame;
extracting images of different levels in a target frame on the infrared image by using a characteristic pyramid algorithm:
and aiming at the images of different levels obtained by the characteristic pyramid, searching a target by utilizing an SSDA template matching algorithm in each layer, selecting the maximum possible position of the target, and simultaneously recording confidence coefficient and target moving position information to obtain the searched target.
2. The method of claim 1, wherein the infrared image is a single-channel 8-bit image, and the image has a pixel value of xinThe image enhancement preprocessing comprises: traversing the image, enhancing the gray value of each pixel, and outputting the pixel value of xout:
Wherein k is the enhancement ratio.
3. The method of claim 1, wherein the extracting of the images of different levels on the infrared image using the feature pyramid algorithm is performing convolution on the target using convolution kernels of 2x2, 3x3, 5x5 and 7x7 to obtain blurred images of different degrees, respectively; different target features are reflected on different feature images, and image information of different dimensions is detected from images of different scales.
4. The method of claim 3, wherein assuming that S (x, y) is a search graph of MxN and T (x, y) is a template graph of MxN, the searching is performed by sliding on the graph to be searched, and the target searching process specifically comprises:
error definition:
wherein the content of the first and second substances,is a subgraph of the search graph, T is a template graph,is the average value of the template graph, the initial position of the upper left corner of the subgraph is (i, j), and then the average subgraphIs as follows;
setting an initial threshold Th0, and when the threshold is matched with a random point in the SSDA algorithm, directly discarding the point if the point is deemed not to be a target by an error accumulation design threshold; the threshold is set to an empirical value;
randomly selecting non-repetitive pixel points in the area to be searched as the center of the tracking frame; calculating the current error, accumulating the error, and recording the current accumulation frequency H when the error accumulation exceeds a threshold Th 0; then traversing all the subgraphs;
in the traversing process, if the error value is greater than the threshold Th0 under the times less than H, the operation of selecting random points to calculate the error is not continued, and the next sub-graph is directly switched to;
in the traversing process, if a subgraph exists, after H times of calculation, the accumulated error is Th 1; if Th1< Th0, update Th0 to Th 1;
recording the matching times H and the accumulated error sum in all the sub-graph matching in the traversal process, and calculating the average error; and after all the subgraphs are traversed, outputting the average error rate and the minimum center coordinate of the subgraph to obtain a search target.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110830767.7A CN113470001B (en) | 2021-07-22 | 2021-07-22 | Target searching method for infrared image |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110830767.7A CN113470001B (en) | 2021-07-22 | 2021-07-22 | Target searching method for infrared image |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113470001A true CN113470001A (en) | 2021-10-01 |
CN113470001B CN113470001B (en) | 2024-01-09 |
Family
ID=77881832
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110830767.7A Active CN113470001B (en) | 2021-07-22 | 2021-07-22 | Target searching method for infrared image |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113470001B (en) |
Citations (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1916538A2 (en) * | 2006-10-27 | 2008-04-30 | Matsushita Electric Works, Ltd. | Target moving object tracking device |
US20140169623A1 (en) * | 2012-12-19 | 2014-06-19 | Microsoft Corporation | Action recognition based on depth maps |
CN103889879A (en) * | 2011-10-19 | 2014-06-25 | 克朗设备公司 | Identifying, matching and tracking multiple objects in a sequence of images |
CN107481210A (en) * | 2017-08-03 | 2017-12-15 | 北京长峰科威光电技术有限公司 | The infrared image enhancing method of local selective mapping based on details |
CN108108657A (en) * | 2017-11-16 | 2018-06-01 | 浙江工业大学 | A kind of amendment local sensitivity Hash vehicle retrieval method based on multitask deep learning |
CN108876723A (en) * | 2018-06-25 | 2018-11-23 | 大连海事大学 | A kind of construction method of the color background of gray scale target image |
CN109754015A (en) * | 2019-01-02 | 2019-05-14 | 京东方科技集团股份有限公司 | Neural network and correlation technique, medium and equipment for the identification of paintings multi-tag |
CN110232419A (en) * | 2019-06-20 | 2019-09-13 | 东北大学 | A kind of method of side slope rock category automatic identification |
CN110363253A (en) * | 2019-07-25 | 2019-10-22 | 安徽工业大学 | A kind of Surfaces of Hot Rolled Strip defect classification method based on convolutional neural networks |
CN111553867A (en) * | 2020-05-15 | 2020-08-18 | 润联软件系统(深圳)有限公司 | Image deblurring method and device, computer equipment and storage medium |
CN111986094A (en) * | 2019-05-22 | 2020-11-24 | 浙江黑卡电气有限公司 | Infrared image optimization processing method based on spatial domain hybrid operation |
CN112184752A (en) * | 2020-09-08 | 2021-01-05 | 北京工业大学 | Video target tracking method based on pyramid convolution |
CN112365973A (en) * | 2020-11-02 | 2021-02-12 | 太原理工大学 | Pulmonary nodule auxiliary diagnosis system based on countermeasure network and fast R-CNN |
CN112560695A (en) * | 2020-12-17 | 2021-03-26 | 中国海洋大学 | Underwater target tracking method, system, storage medium, equipment, terminal and application |
CN112991240A (en) * | 2021-03-19 | 2021-06-18 | 中国科学院长春光学精密机械与物理研究所 | Image self-adaptive enhancement algorithm for real-time image enhancement |
-
2021
- 2021-07-22 CN CN202110830767.7A patent/CN113470001B/en active Active
Patent Citations (15)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP1916538A2 (en) * | 2006-10-27 | 2008-04-30 | Matsushita Electric Works, Ltd. | Target moving object tracking device |
CN103889879A (en) * | 2011-10-19 | 2014-06-25 | 克朗设备公司 | Identifying, matching and tracking multiple objects in a sequence of images |
US20140169623A1 (en) * | 2012-12-19 | 2014-06-19 | Microsoft Corporation | Action recognition based on depth maps |
CN107481210A (en) * | 2017-08-03 | 2017-12-15 | 北京长峰科威光电技术有限公司 | The infrared image enhancing method of local selective mapping based on details |
CN108108657A (en) * | 2017-11-16 | 2018-06-01 | 浙江工业大学 | A kind of amendment local sensitivity Hash vehicle retrieval method based on multitask deep learning |
CN108876723A (en) * | 2018-06-25 | 2018-11-23 | 大连海事大学 | A kind of construction method of the color background of gray scale target image |
CN109754015A (en) * | 2019-01-02 | 2019-05-14 | 京东方科技集团股份有限公司 | Neural network and correlation technique, medium and equipment for the identification of paintings multi-tag |
CN111986094A (en) * | 2019-05-22 | 2020-11-24 | 浙江黑卡电气有限公司 | Infrared image optimization processing method based on spatial domain hybrid operation |
CN110232419A (en) * | 2019-06-20 | 2019-09-13 | 东北大学 | A kind of method of side slope rock category automatic identification |
CN110363253A (en) * | 2019-07-25 | 2019-10-22 | 安徽工业大学 | A kind of Surfaces of Hot Rolled Strip defect classification method based on convolutional neural networks |
CN111553867A (en) * | 2020-05-15 | 2020-08-18 | 润联软件系统(深圳)有限公司 | Image deblurring method and device, computer equipment and storage medium |
CN112184752A (en) * | 2020-09-08 | 2021-01-05 | 北京工业大学 | Video target tracking method based on pyramid convolution |
CN112365973A (en) * | 2020-11-02 | 2021-02-12 | 太原理工大学 | Pulmonary nodule auxiliary diagnosis system based on countermeasure network and fast R-CNN |
CN112560695A (en) * | 2020-12-17 | 2021-03-26 | 中国海洋大学 | Underwater target tracking method, system, storage medium, equipment, terminal and application |
CN112991240A (en) * | 2021-03-19 | 2021-06-18 | 中国科学院长春光学精密机械与物理研究所 | Image self-adaptive enhancement algorithm for real-time image enhancement |
Non-Patent Citations (10)
Title |
---|
F.MEMON 等: "Model based FPGA design of histogram equalization", 《SINDHUNIV. RES. JOUR.》 * |
F.MEMON 等: "Model based FPGA design of histogram equalization", 《SINDHUNIV. RES. JOUR.》, vol. 48, no. 2, 31 December 2016 (2016-12-31), pages 435 - 440 * |
YU WANG 等: "LEDNET: A LIGHTWEIGHT ENCODER-DECODER NETWORK FOR REAL-TIME SEMANTIC SEGMENTATION", ICIP 2019, pages 143 - 144 * |
丁文: "红外图像拼接技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
丁文: "红外图像拼接技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》, vol. 2010, no. 08, 15 August 2010 (2010-08-15), pages 1 - 5 * |
丁文: "红外图像拼接技术研究", 中国优秀硕士学位论文全文数据库 信息科技辑, pages 1 - 5 * |
吴娜娜: "基于机器学习的红外目标跟踪算法研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
吴娜娜: "基于机器学习的红外目标跟踪算法研究", 《中国优秀硕士学位论文全文数据库信息科技辑》, vol. 2020, no. 02, 15 February 2020 (2020-02-15), pages 135 - 87 * |
王利颖 等: "一种自适应红外图像增强新算法", 《信号处理》 * |
王利颖 等: "一种自适应红外图像增强新算法", 《信号处理》, vol. 25, no. 12, 31 December 2009 (2009-12-31), pages 1836 - 1839 * |
Also Published As
Publication number | Publication date |
---|---|
CN113470001B (en) | 2024-01-09 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN111046880B (en) | Infrared target image segmentation method, system, electronic equipment and storage medium | |
CN111741211B (en) | Image display method and apparatus | |
CN112800964B (en) | Remote sensing image target detection method and system based on multi-module fusion | |
JP2015011717A (en) | Ghost artifact detection and removal methods in hdr image processing using multi-scale normalized cross-correlation | |
CN112287868B (en) | Human body action recognition method and device | |
US20190043403A1 (en) | Foreground and background detection method | |
CN109711241B (en) | Object detection method and device and electronic equipment | |
CN113610895A (en) | Target tracking method and device, electronic equipment and readable storage medium | |
CN108764244B (en) | Potential target area detection method based on convolutional neural network and conditional random field | |
CN114627052A (en) | Infrared image air leakage and liquid leakage detection method and system based on deep learning | |
CN112364865B (en) | Method for detecting small moving target in complex scene | |
CN112200035B (en) | Image acquisition method, device and vision processing method for simulating crowded scene | |
CN113780110A (en) | Method and device for detecting weak and small targets in image sequence in real time | |
CN110490196A (en) | Subject detection method and apparatus, electronic equipment, computer readable storage medium | |
CN116486288A (en) | Aerial target counting and detecting method based on lightweight density estimation network | |
CN112435223A (en) | Target detection method, device and storage medium | |
CN116402852A (en) | Dynamic high-speed target tracking method and device based on event camera | |
CN114372931A (en) | Target object blurring method and device, storage medium and electronic equipment | |
CN104809705A (en) | Image denoising method and system based on threshold value block matching | |
CN113724290B (en) | Multi-level template self-adaptive matching target tracking method for infrared image | |
CN112132753A (en) | Infrared image super-resolution method and system for multi-scale structure guide image | |
CN113470001B (en) | Target searching method for infrared image | |
CN113723432B (en) | Intelligent identification and positioning tracking method and system based on deep learning | |
CN115171011A (en) | Multi-class building material video counting method and system and counting equipment | |
CN113240611A (en) | Foreign matter detection method based on picture sequence |
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 |