CN111583279A - Super-pixel image segmentation method based on PCBA - Google Patents
Super-pixel image segmentation method based on PCBA Download PDFInfo
- Publication number
- CN111583279A CN111583279A CN202010397755.5A CN202010397755A CN111583279A CN 111583279 A CN111583279 A CN 111583279A CN 202010397755 A CN202010397755 A CN 202010397755A CN 111583279 A CN111583279 A CN 111583279A
- Authority
- CN
- China
- Prior art keywords
- image
- segmentation
- superpixel
- value
- pixel
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 29
- RVCKCEDKBVEEHL-UHFFFAOYSA-N 2,3,4,5,6-pentachlorobenzyl alcohol Chemical compound OCC1=C(Cl)C(Cl)=C(Cl)C(Cl)=C1Cl RVCKCEDKBVEEHL-UHFFFAOYSA-N 0.000 title claims abstract description 21
- 238000003709 image segmentation Methods 0.000 title claims abstract description 18
- 230000011218 segmentation Effects 0.000 claims abstract description 75
- 239000003086 colorant Substances 0.000 claims description 27
- 238000001914 filtration Methods 0.000 claims description 12
- 238000004364 calculation method Methods 0.000 abstract description 9
- 230000000007 visual effect Effects 0.000 description 5
- 238000000605 extraction Methods 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 238000003064 k means clustering Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 230000009466 transformation 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
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- 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
- G06T2207/10004—Still image; Photographic image
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a PCBA-based super-pixel image segmentation method, which comprises the following steps: acquiring an original image; improving the contrast of the original image to obtain an input image; calculating a saliency value of an input image and generating a saliency map; performing superpixel segmentation on the saliency map, and acquiring a superpixel label; creating a super-pixel image based on the super-pixel label, wherein the pixel value of each pixel in the super-pixel image is the average gray value of the corresponding super-pixel block; and segmenting the original image into a target region and a background region based on the optimal segmentation threshold value, and generating a segmentation map consisting of the target region and the background region. Aiming at the characteristics that the color information of the PCBA is simple and does not contain complex background information, the invention generates the superpixel image by using the superpixel label before the final segmentation, thus not only retaining important information such as the edge of the original image and the like, but also reducing the pixels of the image needing to be processed, further reducing the calculation cost and improving the segmentation efficiency.
Description
Technical Field
The invention belongs to the field of image processing, and particularly relates to a super-pixel image segmentation method based on PCBA.
Background
A super pixel belongs to an image segmentation technology and refers to an irregular pixel block which is formed by adjacent pixels with similar texture, color, brightness and other characteristics and has a certain visual significance. The superpixel segmentation algorithm is an important preprocessing tool for reducing calculation, and is widely applied to the fields of target tracking, target identification, 3D reconstruction, image segmentation, salient feature extraction and the like. The super-pixels can remove redundant information relative to the pixels, so that the operation speed is improved, the super-pixels can be used as a common means for preprocessing in significance detection, and the operation efficiency can be greatly improved while the edge information is saved as much as possible. Currently, superpixel algorithms can be simply classified into gradient descent-based algorithms and clustering-based algorithms.
Both the watershed algorithm and the Mean-Shift algorithm are classical gradient descent algorithms. The watershed algorithm uses gradient descent to obtain the similarity between pixels to generate a closed contour. The Mean-Shift algorithm is an iterative process that finds the average motion vector of the current pixel, moves to the vector position, and continues the search until a certain condition is met. Although the watershed algorithm has a faster operation speed, the segmentation is easy to carry out, and Mean-Shift is opposite. The K-means Clustering (K-means) algorithm, the Linear Spectral Clustering (LSC) algorithm, the Simple Linear Iterative Clustering (SLIC) algorithm and the Simple Non-Iterative Clustering (Simple Non-Iterative Clustering) algorithm are all classical Clustering algorithms. The LSC algorithm, the SLIC algorithm and the SNIC algorithm are all improvements based on the K-means algorithm. The K-means algorithm randomly selects K targets as initial clustering centers, then calculates the distance from each target to the initial centers, and classifies each target to a nearest central point. The K-means algorithm requires an initialization center and the number of iterations depends on the input picture. The SNIC algorithm does not need iteration, only needs little memory, and has a speed superior to that of the SLIC algorithm and the LSC algorithm.
Although the above algorithm has a good segmentation result for natural images, some objects are missed when segmenting PCBA images.
Disclosure of Invention
Aiming at the defects in the prior art, the problems to be solved by the invention are as follows: how to avoid missing targets when segmenting PCBA images.
In order to solve the technical problems, the invention adopts the following technical scheme:
a super pixel image segmentation method based on PCBA includes:
s1, acquiring an original image;
s2, improving the contrast of the original image to obtain an input image;
s3, quantizing the RGB color channels of the input image into m different colors, selecting n colors with the highest occurrence frequency in the input image as high-frequency colors to represent all input colors, calculating the saliency value of the input image and generating a saliency map;
s4, performing superpixel segmentation on the saliency map, and acquiring a superpixel label;
s5, creating a super pixel image based on the super pixel label, wherein the pixel value of each pixel in the super pixel image is the average gray value of the corresponding super pixel block;
s6, obtaining an optimal segmentation threshold value by adopting a maximum threshold segmentation method for the super-pixel image, segmenting the original image into a target area and a background area based on the optimal segmentation threshold value, and generating a segmentation image consisting of the target area and the background area.
Preferably, the formula for calculating the significant value in step S3 is as follows:
wherein S (c) is the significant value of the input color c, ciI-th nearest neighbor color of the input color c, k is the number of nearest neighbor colors, S (c)i) Is ciThe corresponding initial value of the significance is,
Pjprobability of the jth nearest neighbor color of input color c, D (c)i,cj) I nearest neighbor color of input color c and j nearest neighbor color in LAB colorThe distance in space, D is the sum of the distances of the input color c to its k nearest neighbors,
preferably, in step S4, the saliency map is super-pixel segmented using an SNIC super-pixel segmentation algorithm.
Preferably, step S6 includes:
s601, calculating an initial segmentation threshold which is equal to the average value of the maximum value and the minimum value of the gray value of the super pixel image;
s602, segmenting the superpixel image into a superpixel target region and a superpixel background region based on the initial segmentation threshold, and iteratively updating the initial segmentation threshold through an information entropy formula to obtain an optimal segmentation threshold;
s603, segmenting the original image into a target region and a background region based on the optimal segmentation threshold;
s604, generating a segmentation map consisting of the target area and the background area.
Preferably, the method further comprises the following steps:
and S7, filtering the region of the segmentation graph with the target area smaller than the filtering threshold value into a background region.
In summary, the present invention discloses a method for segmenting a superpixel image based on PCBA, comprising: s1, acquiring an original image; s2, improving the contrast of the original image to obtain an input image; s3, quantizing the RGB color channels of the input image into m different colors, selecting n colors with the highest occurrence frequency in the input image as high-frequency colors to represent all input colors, calculating the saliency value of the input image and generating a saliency map; s4, performing superpixel segmentation on the saliency map, and acquiring a superpixel label; s5, creating a super pixel image based on the super pixel label, wherein the pixel value of each pixel in the super pixel image is the average gray value of the corresponding super pixel block; s6, obtaining an optimal segmentation threshold value by adopting a maximum threshold segmentation method for the super-pixel image, segmenting the original image into a target area and a background area based on the optimal segmentation threshold value, and generating a segmentation image consisting of the target area and the background area.
Compared with the existing superpixel segmentation algorithm, the method has the following beneficial effects:
the invention firstly provides a super-pixel label image method to improve the efficiency of image segmentation. Meanwhile, the visual system is sensitive to the contrast of visual signals, the effect of extracting the significant features based on the histogram contrast is obvious, the SNIC superpixel segmentation algorithm utilizes color and space information to perform primary segmentation of images, iteration is not needed although the method is a clustering method, and the speed is superior to that of a general clustering algorithm, so that the histogram contrast is adopted to perform significant feature extraction, and then the superpixel segmentation is performed.
Drawings
FIG. 1 is a flow chart of one embodiment of a PCBA-based superpixel image segmentation method disclosed in the present invention;
FIG. 2 is a saliency map illustration of the present invention;
FIG. 3 is an example of a superpixel segmentation map in the present invention;
FIG. 4 is a schematic diagram of a superpixel image generation process;
FIG. 5 is a graph of segmentation results for different optimal segmentation thresholds;
FIG. 6 is a diagram of the final segmentation result after the area of the filtering target region is less than 70;
FIG. 7 is a comparison of saliency maps for different algorithms, with line (a) being the original picture, line (b) being the group Truth, and lines (c) - (h) being the results maps for Watershed, SNIC, Mean-Shift, LSC, SLIC, and K-means algorithms, respectively, and (i) being the results map for the present invention;
FIG. 8 is a comparison graph of PR curves for different algorithms;
FIG. 9 is a graph comparing ROC curves for different algorithms.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings.
As shown in FIG. 1, the invention discloses a super-pixel image segmentation method based on PCBA, comprising the following steps:
s1, acquiring an original image;
because the initial PCBA picture was taken by the industrial camera with average pixels around 2777 x 2138, the original image can be cropped for subsequent calculations, e.g., to a size of 512 x 512 in common image processing size.
S2, improving the contrast of the original image to obtain an input image;
in order to facilitate subsequent feature extraction, image preprocessing can be performed on an original image, a stretchlim function in Matlab software which can be directly used can adaptively find a segmentation threshold vector to change the contrast of an image, and a gray level transformation adaptive optimal threshold can be obtained.
S3, quantizing the RGB color channels of the input image into m different colors, selecting n colors with the highest occurrence frequency in the input image as high-frequency colors to represent all input colors, calculating the saliency value of the input image and generating a saliency map;
specifically, 12 different colors, 85 high frequency colors are selected to represent all input colors. The visual system is sensitive to the contrast of the visual signal, and therefore, it is obvious to extract the salient features based on the histogram contrast.
S4, performing superpixel segmentation on the saliency map, and acquiring a superpixel label;
s5, creating a super pixel image based on the super pixel label, wherein the pixel value of each pixel in the super pixel image is the average gray value of the corresponding super pixel block;
as shown in fig. 4, for simplicity of calculation, the present invention proposes a method for generating a new super-pixel image by using label information, where each pixel value of the super-pixel image is an average gray value of a super-pixel block, and a calculation formula is as follows:
I(i)＝f(O(i)),i＝1,2,…,L
where f is the mean function, X and Y are the coordinate set of label ═ i, g (X, Y) is the pixel value of (X, Y), and o (i) is the gray value of the pixel block corresponding to the ith label, i.e. the sum of the gray values of all the pixels belonging to the ith label. These new pixel values are arranged in a 28 x 28 matrix from the original position left to right to form a new graph.
In the prior art, an original image is generally directly operated, the image is segmented by acquiring the information entropy of the original image or the like, or the image is directly segmented by using an SNIC (simple network integration) algorithm, and a superpixel image generated by using a superpixel label is not seen at present, because the size of a pixel of the superpixel image is small, but important information such as the edge of the original image is reserved, the computing speed is greatly improved, the speed for computing the information entropy by using the method is higher, and the influence of information loss is small.
S6, obtaining an optimal segmentation threshold value by adopting a maximum threshold segmentation method for the super-pixel image, segmenting the original image into a target area and a background area based on the optimal segmentation threshold value, and generating a segmentation image consisting of the target area and the background area.
Compared with the prior art, aiming at the characteristics that the color information of the PCBA is simple and does not contain complex background information, the method generates the superpixel image by using the superpixel label before the final segmentation, thus not only retaining important information such as the edge of the original image and the like, but also not causing target omission, reducing the pixels of the image needing to be processed, further reducing the calculation cost and improving the segmentation efficiency.
As shown in fig. 2, in the implementation, the formula for calculating the significant value in step S3 is as follows:
wherein S (c) is the significant value of the input color c, ciFor the ith nearest neighbor color of the input color c,k is the number of nearest neighbor colors, S (c)i) Is ciThe corresponding initial value of the significance is,
Pjis the jth nearest neighbor color c of the input color cjI.e. the gray value is the jth nearest neighbor color cjThe proportion of the number of pixels of the corresponding gray value to the total number of pixels of the input image, D (c)i,cj) The i-th nearest neighbor color of the input color c and the j-th nearest neighbor color are within the LAB color space, D is the sum of the distances of the input color c from its k nearest neighbor colors,
the RGB color channels can be quantized to obtain 12 different values, high-frequency colors are selected, the selected colors cover more than 95% of pixels, and the number of the high-frequency colors can be 85.
The color coverage of more than 95 percent means that the selected color pixel value accounts for more than 95 percent of the color pixel value of the whole image, and the calculation amount can be reduced by selecting the color pixel value to replace the color of the whole image.
In the specific implementation, as shown in fig. 3, in step S4, the saliency map is super-pixel divided using the SNIC super-pixel division algorithm.
The super-pixel segmentation of the salient image can adopt SLIC algorithm and SNIC algorithm, the obtained salient image is subjected to super-pixel segmentation by using SNIC algorithm, the SNIC algorithm recalculates the average value of each super-pixel block and replaces the central value of the pixel block when a new pixel is added into the super-pixel block each time, iteration is not needed, and the speed is superior to that of the common clustering algorithm. In addition, the search range of the SNIC algorithm is 4 or 8 neighborhoods of each central point, so that the phenomenon that some pixels are mistakenly divided into the superpixel blocks due to the fact that the distance between the pixels and the central points of the superpixel blocks is short can be avoided, and the dividing precision is improved.
The super-pixel distance calculation formula is as follows:
wherein d islabIs the color distance of the LAB color space, dxyIs the color distance between two pixels, l, a and b are the 3 color feature components of the LAB color space, respectively, x and y are the coordinate position components,
is a balance factor, N is the number of input image pixels, L is the number of superpixels, and M is a balance factor. Specifically, the balance factor M is 20, and the number L of super pixels is 800.
As shown in fig. 5, in a specific implementation, step S6 includes:
s601, calculating an initial segmentation threshold which is equal to the average value of the maximum value and the minimum value of the gray value of the super pixel image;
s602, segmenting the superpixel image into a superpixel target region and a superpixel background region based on the initial segmentation threshold, and iteratively updating the initial segmentation threshold through an information entropy formula to obtain an optimal segmentation threshold;
s603, segmenting the original image into a target region and a background region based on the optimal segmentation threshold;
s604, generating a segmentation map consisting of the target area and the background area.
The information entropy calculation formula is as follows:
wherein HoAnd HbEntropy of information, H, for the target region and the background region, respectivelyIIs the information entropy of the picture when HIObtaining an optimal segmentation threshold when a maximum value is reached after iteration, (q, r) being the optimal segmentation threshold TmaxIs determined by the coordinate of (a) in the space,
and
is the sum of the total probabilities of the target region and the background region, respectively, and Pb＝1-PoAnd p (i, j) is the probability of the pixel point (i, j) in the histogram. And performing threshold segmentation on the original image according to the obtained optimal segmentation threshold to obtain a segmentation image.
As shown in fig. 6, in specific implementation, the method further includes:
and S7, filtering the region of the segmentation graph with the target area smaller than the filtering threshold value into a background region.
For the effect of the super-pixel image segmentation method based on PCBA disclosed by the invention, the following performance reference indexes can be used for evaluation:
wherein S represents a normalization to [0,1 ]]The better the algorithm performance, G represents the true significance calibration, W and H are the width and height of the corresponding image, the smaller the MAE value is, TP represents the number of pixels in the binary significance result that are consistent between the target region and the true significance calibration, TN represents the number of pixels in the binary significance result that are consistent between the background region and the true significance calibration, FP represents the number of pixels that are incorrectly divided into the target in the binary significance result, FN represents the number of pixels that are incorrectly divided into the background in the binary significance result, Precision is the Precision, which is the ratio of the correctly calibrated pixels among all foreground pixels generated by the algorithm, Recall is the Recall, which is the ratio of the correctly calibrated pixels by the algorithm in the foreground pixels actually calibrated by the true value, with the Recall being the horizontal axis and the Precision being the vertical axis, a PR curve can be obtained, the closer to the upper right, β is the better the algorithm performance2＝0.3，Fβ(F-measure) simultaneously considering precision ratio and recall ratio, algorithm performance and FβThe values of the parameters are in direct proportion, FPR is False Positive probability (False Positive Rate), the horizontal axis is drawn, TPR is True Positive probability (True Positive Rate), the vertical axis is drawn, the closer the drawn ROC Curve is to the upper left corner, the better the algorithm performance is represented, and the AUC index (the Area Under the receiver operating characteristics Curve), namely the Area Under the ROC Curve, the better the algorithm performance is represented.
Taking the original image size as 512 × 512, quantizing the RGB color channels to obtain 12 different values, taking the high-frequency color number as 85, where M is 20 and L is 800 as an example, the label number is 784, and the super-pixel picture size is 28 × 28. Table 1 compares MAE values at different filtering thresholds against run time.
TABLE 1 evaluation index comparison table under different filtering threshold values
As can be seen from table 1, when the filtering threshold is 70, the value of MAE tends to be smooth, and the operation time also tends to be smooth, so 70 is selected as the filtering threshold of the present invention.
After image segmentation is completed, a saliency map, an MAE (maximum intensity enhancement), an F-measure, a PR (particle-correlation) curve, an ROC (rock-angle) curve and running time in an experiment are obtained through statistics and serve as effective model performance reference indexes, and a table 2 shows comparison between the MAE, the F-measure and the running time under different algorithms.
TABLE 2 quantitative comparison table under different algorithms
As can be seen from Table 2, the present invention has the highest F-measure value and the lowest MAE value, and although the running time is lower than that of the Watershed algorithm, other indexes are far better than that of the Watershed algorithm.
As shown in fig. 7 to 9, it can be seen from the above experiments that, aiming at the phenomenon of some under-segmentation existing when the conventional image superpixel segmentation algorithm segments the PCBA image, a superpixel image segmentation algorithm based on the PCBA is provided, and meanwhile, the algorithm efficiency and the accuracy are considered. The method comprises the steps of extracting significance characteristics by using a gradient histogram contrast algorithm, segmenting a significance map by using a superpixel segmentation algorithm, generating a new image by using a superpixel label obtained by the segmentation algorithm, extracting the information entropy of the new image, segmenting an original image by using an information entropy threshold, filtering out regions with the target area smaller than 70 in the segmentation map, and obtaining a final segmentation result map, wherein the final F-measure value is 0.7979, the MAE value is 0.040567, and the segmentation result map is superior to other six algorithms.
The above is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several changes and modifications can be made without departing from the technical solution, and the technical solution of the changes and modifications should be considered as falling within the scope of the claims of the present application.
Claims (5)
1. A super pixel image segmentation method based on PCBA is characterized by comprising the following steps:
s1, acquiring an original image;
s2, improving the contrast of the original image to obtain an input image;
s3, quantizing the RGB color channels of the input image into m different colors, selecting n colors with the highest occurrence frequency in the input image as high-frequency colors to represent all input colors, calculating the saliency value of the input image and generating a saliency map;
s4, performing superpixel segmentation on the saliency map, and acquiring a superpixel label;
s5, creating a super pixel image based on the super pixel label, wherein the pixel value of each pixel in the super pixel image is the average gray value of the corresponding super pixel block;
s6, obtaining an optimal segmentation threshold value by adopting a maximum threshold segmentation method for the super-pixel image, segmenting the original image into a target area and a background area based on the optimal segmentation threshold value, and generating a segmentation image consisting of the target area and the background area.
2. The PCBA-based superpixel image segmentation method according to claim 1, wherein the formula for calculating the saliency value in step S3 is as follows:
wherein S (c) is the significant value of the input color c, ciI-th nearest neighbor color of the input color c, k is the number of nearest neighbor colors, S (c)i) Is ciThe corresponding initial value of the significance is,
Pjprobability of the jth nearest neighbor color of input color c, D (c)i,cj) The i-th nearest neighbor color of the input color c and the j-th nearest neighbor color are within the LAB color space, D is the sum of the distances of the input color c from its k nearest neighbor colors,
3. the PCBA-based superpixel image segmentation method of claim 1, wherein in step S4, said saliency map is superpixel segmented using a SNIC superpixel segmentation algorithm.
4. The PCBA-based superpixel image segmentation method of claim 1, wherein step S6 comprises:
s601, calculating an initial segmentation threshold which is equal to the average value of the maximum value and the minimum value of the gray value of the super pixel image;
s602, segmenting the superpixel image into a superpixel target region and a superpixel background region based on the initial segmentation threshold, and iteratively updating the initial segmentation threshold through an information entropy formula to obtain an optimal segmentation threshold;
s603, segmenting the original image into a target region and a background region based on the optimal segmentation threshold;
s604, generating a segmentation map consisting of the target area and the background area.
5. The PCBA-based superpixel image segmentation method of claim 1, further comprising:
and S7, filtering the region of the segmentation graph with the target area smaller than the filtering threshold value into a background region.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010397755.5A CN111583279A (en) | 2020-05-12 | 2020-05-12 | Super-pixel image segmentation method based on PCBA |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010397755.5A CN111583279A (en) | 2020-05-12 | 2020-05-12 | Super-pixel image segmentation method based on PCBA |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111583279A true CN111583279A (en) | 2020-08-25 |
Family
ID=72112172
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010397755.5A Pending CN111583279A (en) | 2020-05-12 | 2020-05-12 | Super-pixel image segmentation method based on PCBA |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111583279A (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108447050A (en) * | 2018-03-07 | 2018-08-24 | 湘潭大学 | A kind of Surface Flaw dividing method based on super-pixel |
| CN112085749A (en) * | 2020-09-10 | 2020-12-15 | 桂林电子科技大学 | Multi-scale non-iterative superpixel segmentation method |
| CN113205525A (en) * | 2021-03-23 | 2021-08-03 | 中国科学院南京土壤研究所 | Soil image segmentation and extraction method based on super-green index and super-pixel |
| CN113409338A (en) * | 2021-06-24 | 2021-09-17 | 西安交通大学 | Super-pixel method based on probability distribution |
| CN113506266A (en) * | 2021-07-09 | 2021-10-15 | 平安科技（深圳）有限公司 | Method, device and equipment for detecting tongue greasy coating and storage medium |
| CN113593008A (en) * | 2021-07-06 | 2021-11-02 | 四川大学 | True 3D image significant reconstruction method under complex scene |
| CN113808206A (en) * | 2021-09-14 | 2021-12-17 | 广东三水合肥工业大学研究院 | Typesetting system and method based on visual tracking robot |
| CN114049360A (en) * | 2022-01-13 | 2022-02-15 | 南通海恒纺织设备有限公司 | Textile dyeing toner mixing control method and system based on graph cut algorithm |
Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105261037A (en) * | 2015-10-08 | 2016-01-20 | 重庆理工大学 | Moving object detection method capable of automatically adapting to complex scenes |
| CN105809651A (en) * | 2014-12-16 | 2016-07-27 | 吉林大学 | Image saliency detection method based on edge non-similarity comparison |
| CN106296695A (en) * | 2016-08-12 | 2017-01-04 | 西安理工大学 | Adaptive threshold natural target image based on significance segmentation extraction algorithm |
| CN106447679A (en) * | 2016-10-17 | 2017-02-22 | 大连理工大学 | Obviousness detection method based on grabcut and adaptive cluster clustering |
| CN106981068A (en) * | 2017-04-05 | 2017-07-25 | 重庆理工大学 | A kind of interactive image segmentation method of joint pixel pait and super-pixel |
| CN107229917A (en) * | 2017-05-31 | 2017-10-03 | 北京师范大学 | A kind of several remote sensing image general character well-marked target detection methods clustered based on iteration |
| CN107967480A (en) * | 2016-10-19 | 2018-04-27 | 北京联合大学 | A kind of notable object extraction method based on label semanteme |
| CN108921820A (en) * | 2018-05-30 | 2018-11-30 | 咸阳师范学院 | A kind of saliency object detection method based on feature clustering and color contrast |
| CN109636784A (en) * | 2018-12-06 | 2019-04-16 | 西安电子科技大学 | Saliency object detection method based on maximum neighborhood and super-pixel segmentation |
| CN109685806A (en) * | 2018-11-14 | 2019-04-26 | 武汉科技大学 | Image significance detection method and device |
| CN109829423A (en) * | 2019-01-18 | 2019-05-31 | 电子科技大学 | A kind of icing lake infrared imaging detection method |
| CN109886267A (en) * | 2019-01-29 | 2019-06-14 | 杭州电子科技大学 | A kind of soft image conspicuousness detection method based on optimal feature selection |
| CN110428371A (en) * | 2019-07-03 | 2019-11-08 | 深圳大学 | Image defogging method, system, storage medium and electronic equipment based on super-pixel segmentation |
| CN110866896A (en) * | 2019-10-29 | 2020-03-06 | 中国地质大学（武汉） | Image saliency target detection method based on k-means and level set super-pixel segmentation |
-
2020
- 2020-05-12 CN CN202010397755.5A patent/CN111583279A/en active Pending
Patent Citations (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105809651A (en) * | 2014-12-16 | 2016-07-27 | 吉林大学 | Image saliency detection method based on edge non-similarity comparison |
| CN105261037A (en) * | 2015-10-08 | 2016-01-20 | 重庆理工大学 | Moving object detection method capable of automatically adapting to complex scenes |
| CN106296695A (en) * | 2016-08-12 | 2017-01-04 | 西安理工大学 | Adaptive threshold natural target image based on significance segmentation extraction algorithm |
| CN106447679A (en) * | 2016-10-17 | 2017-02-22 | 大连理工大学 | Obviousness detection method based on grabcut and adaptive cluster clustering |
| CN107967480A (en) * | 2016-10-19 | 2018-04-27 | 北京联合大学 | A kind of notable object extraction method based on label semanteme |
| CN106981068A (en) * | 2017-04-05 | 2017-07-25 | 重庆理工大学 | A kind of interactive image segmentation method of joint pixel pait and super-pixel |
| CN107229917A (en) * | 2017-05-31 | 2017-10-03 | 北京师范大学 | A kind of several remote sensing image general character well-marked target detection methods clustered based on iteration |
| CN108921820A (en) * | 2018-05-30 | 2018-11-30 | 咸阳师范学院 | A kind of saliency object detection method based on feature clustering and color contrast |
| CN109685806A (en) * | 2018-11-14 | 2019-04-26 | 武汉科技大学 | Image significance detection method and device |
| CN109636784A (en) * | 2018-12-06 | 2019-04-16 | 西安电子科技大学 | Saliency object detection method based on maximum neighborhood and super-pixel segmentation |
| CN109829423A (en) * | 2019-01-18 | 2019-05-31 | 电子科技大学 | A kind of icing lake infrared imaging detection method |
| CN109886267A (en) * | 2019-01-29 | 2019-06-14 | 杭州电子科技大学 | A kind of soft image conspicuousness detection method based on optimal feature selection |
| CN110428371A (en) * | 2019-07-03 | 2019-11-08 | 深圳大学 | Image defogging method, system, storage medium and electronic equipment based on super-pixel segmentation |
| CN110866896A (en) * | 2019-10-29 | 2020-03-06 | 中国地质大学（武汉） | Image saliency target detection method based on k-means and level set super-pixel segmentation |
Non-Patent Citations (3)
| Title |
|---|
| ACHANTA R等: "Superpixels and polygons using simple non-iterative clustering", 《CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION》, pages 19 - 23 * |
| 廖苗 等: "一种新的图像超像素分割方法", 《电子与信息学报》 * |
| 李然 等 * |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108447050A (en) * | 2018-03-07 | 2018-08-24 | 湘潭大学 | A kind of Surface Flaw dividing method based on super-pixel |
| CN112085749A (en) * | 2020-09-10 | 2020-12-15 | 桂林电子科技大学 | Multi-scale non-iterative superpixel segmentation method |
| CN112085749B (en) * | 2020-09-10 | 2022-07-05 | 桂林电子科技大学 | Multi-scale non-iterative superpixel segmentation method |
| CN113205525A (en) * | 2021-03-23 | 2021-08-03 | 中国科学院南京土壤研究所 | Soil image segmentation and extraction method based on super-green index and super-pixel |
| CN113409338A (en) * | 2021-06-24 | 2021-09-17 | 西安交通大学 | Super-pixel method based on probability distribution |
| CN113593008A (en) * | 2021-07-06 | 2021-11-02 | 四川大学 | True 3D image significant reconstruction method under complex scene |
| CN113506266A (en) * | 2021-07-09 | 2021-10-15 | 平安科技（深圳）有限公司 | Method, device and equipment for detecting tongue greasy coating and storage medium |
| CN113808206A (en) * | 2021-09-14 | 2021-12-17 | 广东三水合肥工业大学研究院 | Typesetting system and method based on visual tracking robot |
| CN113808206B (en) * | 2021-09-14 | 2024-02-02 | 广东三水合肥工业大学研究院 | Typesetting system and method based on vision tracking robot |
| CN114049360A (en) * | 2022-01-13 | 2022-02-15 | 南通海恒纺织设备有限公司 | Textile dyeing toner mixing control method and system based on graph cut algorithm |
| CN114049360B (en) * | 2022-01-13 | 2022-03-22 | 南通海恒纺织设备有限公司 | Textile dyeing toner mixing control method and system based on graph cut algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111583279A (en) | Super-pixel image segmentation method based on PCBA | |
| CN112424828B (en) | Nuclear fuzzy C-means quick clustering algorithm integrating space constraint | |
| CN110866896B (en) | Image saliency target detection method based on k-means and level set super-pixel segmentation | |
| CN108537239B (en) | Method for detecting image saliency target | |
| CN110111338B (en) | Visual tracking method based on superpixel space-time saliency segmentation | |
| CN108629783B (en) | Image segmentation method, system and medium based on image feature density peak search | |
| CN106815842B (en) | improved super-pixel-based image saliency detection method | |
| CN111340824B (en) | Image feature segmentation method based on data mining | |
| CN109086777B (en) | Saliency map refining method based on global pixel characteristics | |
| CN111339924B (en) | Polarized SAR image classification method based on superpixel and full convolution network | |
| CN114492619B (en) | Point cloud data set construction method and device based on statistics and concave-convex performance | |
| CN113888461A (en) | Method, system and equipment for detecting defects of hardware parts based on deep learning | |
| CN115471682A (en) | Image matching method based on SIFT fusion ResNet50 | |
| CN112329784A (en) | Correlation filtering tracking method based on space-time perception and multimodal response | |
| CN110782487A (en) | Target tracking method based on improved particle filter algorithm | |
| CN109741358B (en) | Superpixel segmentation method based on adaptive hypergraph learning | |
| CN113657225B (en) | Target detection method | |
| CN111091129A (en) | Image salient region extraction method based on multi-color characteristic manifold sorting | |
| CN111274964A (en) | Detection method for analyzing water surface pollutants based on visual saliency of unmanned aerial vehicle | |
| CN113627481A (en) | Multi-model combined unmanned aerial vehicle garbage classification method for smart gardens | |
| CN111539966A (en) | Colorimetric sensor array image segmentation method based on fuzzy c-means clustering | |
| CN108573217B (en) | Compression tracking method combined with local structured information | |
| CN107085725B (en) | Method for clustering image areas through LLC based on self-adaptive codebook | |
| CN114549649A (en) | Feature matching-based rapid identification method for scanned map point symbols | |
| CN110490877B (en) | Target segmentation method for binocular stereo image based on Graph Cuts |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200825 |