CN115953335A - Image median filtering method, system, medium and electronic device - Google Patents
Image median filtering method, system, medium and electronic device Download PDFInfo
- Publication number
- CN115953335A CN115953335A CN202310238322.9A CN202310238322A CN115953335A CN 115953335 A CN115953335 A CN 115953335A CN 202310238322 A CN202310238322 A CN 202310238322A CN 115953335 A CN115953335 A CN 115953335A
- Authority
- CN
- China
- Prior art keywords
- gray value
- median
- sequence
- value
- gray
- 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
- 238000001914 filtration Methods 0.000 title claims abstract description 51
- 238000000034 method Methods 0.000 title claims abstract description 35
- 239000011159 matrix material Substances 0.000 claims abstract description 26
- 238000012163 sequencing technique Methods 0.000 claims abstract description 15
- 239000013598 vector Substances 0.000 description 5
- 238000010586 diagram Methods 0.000 description 3
- 238000002864 sequence alignment Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 235000002566 Capsicum Nutrition 0.000 description 1
- 239000006002 Pepper Substances 0.000 description 1
- 235000016761 Piper aduncum Nutrition 0.000 description 1
- 235000017804 Piper guineense Nutrition 0.000 description 1
- 244000203593 Piper nigrum Species 0.000 description 1
- 235000008184 Piper nigrum Nutrition 0.000 description 1
- 238000010009 beating Methods 0.000 description 1
- 239000000872 buffer Substances 0.000 description 1
- 238000007796 conventional method Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 150000003839 salts Chemical class 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
Images
Classifications
-
- 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
Abstract
The invention provides an image median filtering method, system, medium and electronic equipment, and belongs to the technical field of image processing. The method comprises the following steps: acquiring a pixel matrix taking a certain pixel point as a center, and sequencing each one-dimensional sequence of the pixel matrix according to the gray value to obtain the gray value median of each one-dimensional sequence; sorting the gray value median values of all the one-dimensional sequences according to the size, and taking the median value of the gray value median sorting result as an update value of the gray value of a central pixel point of the pixel matrix; and sequentially carrying out the processing on each pixel point of the image to obtain a median filtering result of the image. The invention realizes the quick search of the median of the gray values in the current window and greatly improves the rate of image median filtering processing.
Description
Technical Field
The present invention relates to the field of image processing technologies, and in particular, to a method, a system, a medium, and an electronic device for median filtering of an image.
Background
The statements in this section merely provide background information related to the present disclosure and may not constitute prior art.
The sampling or transmission of digital images is often disturbed by noise when passing through a sensor or transmission channel, and image filtering is an essential step in image preprocessing, which not only preserves image detail characteristics to the maximum extent but also removes noise information in the image.
The median filtering is a typical nonlinear filtering method, the filtering effect is significant under the condition that the noise is shot noise, the median filtering is a filtering method based on a sequencing statistical theory, and the basic idea is as follows: the gray values of all pixel points in a sliding window taking a certain pixel as the center are sequenced, an intermediate value is selected as a new value of the gray value of the center pixel point, and the median filtering effect is particularly ideal for speckle noise and salt and pepper noise.
However, data sorting in the general median filtering is time-consuming, and especially under a large sliding window, a large amount of data comparison work needs to be performed, which is not beneficial to fast and real-time processing of images.
Disclosure of Invention
In order to solve the defects of the prior art, the invention provides an image median filtering method, a system, a medium and electronic equipment, which realize the quick search of the median of the gray values in the current window and greatly improve the rate of image median filtering processing.
In order to achieve the purpose, the invention adopts the following technical scheme:
the invention provides a median filtering method for an image in a first aspect.
An image median filtering method, comprising the processes of:
acquiring a pixel matrix taking a certain pixel point as a center, and sequencing each one-dimensional sequence of the pixel matrix according to the gray value to obtain the gray value median of each one-dimensional sequence; sorting the gray value median values of all the one-dimensional sequences according to the size, and taking the median value of the gray value median sorting result as an update value of the gray value of a central pixel point of the pixel matrix;
and sequentially carrying out the processing on each pixel point of the image to obtain a median filtering result of the image.
As a further limitation of the first aspect of the present invention, acquiring a pixel matrix centered on a certain pixel point, and sorting the one-dimensional sequences of the pixel matrix according to the gray value size to obtain the gray value median of each one-dimensional sequence, includes:
configuring continuous serial numbers for all gray values of any one-dimensional sequence according to the sequence of entering a processing flow, and recording the result as 1 when the current gray value is larger than the gray value before the corresponding sequence of the gray value, or recording the result as 0; when the current gray value is greater than or equal to the gray value after the sequence corresponding to the gray value, recording the result as 1, otherwise, recording the result as 0;
comparing each gray value of the current one-dimensional sequence with other gray values of the current one-dimensional sequence in sequence to obtain a comparison and addition result corresponding to each gray value, and obtaining a gray value median of the current one-dimensional sequence by taking the comparison and addition result corresponding to each gray value as a sequencing serial number;
and performing the sorting on the gray values of the one-dimensional sequences to obtain the gray value median of the one-dimensional sequences.
As a further limitation of the first aspect of the present invention, the sorting of the median of the gray values of the respective one-dimensional sequences by size comprises:
sorting the grey value median values, recording the result as 1 when the current grey value median value is greater than the grey value median value before the corresponding sequence of the grey value median value, and recording the result as 0 if the current grey value median value is not greater than the grey value median value before the corresponding sequence of the grey value median value; when the current gray value median is greater than or equal to the gray value median after the order corresponding to the gray value median, recording the result as 1, otherwise, recording the result as 0;
and comparing the gray value median with other gray value medias in sequence to obtain a comparison and addition result corresponding to each gray value median, taking the comparison and addition result corresponding to each gray value median as a sorting serial number to obtain the median of each gray value median, and taking the median of each gray value median as an updated value of the gray value of the central pixel point of the sliding window.
As a further limitation of the first aspect of the invention, the one-dimensional sequence is a sequence of rows or a sequence of columns.
The invention provides a median filtering system for an image in a second aspect.
An image median filtering system, comprising:
a grayscale ranking module configured to: acquiring a pixel matrix taking a certain pixel point as a center, and sequencing each one-dimensional sequence of the pixel matrix according to the gray value to obtain the gray value median of each one-dimensional sequence; sorting the gray value median values of all the one-dimensional sequences according to the size, and taking the median value of the gray value median sorting result as an update value of the gray value of a central pixel point of the pixel matrix;
a median filtering module configured to: and sequentially carrying out the processing on each pixel point of the image to obtain a median filtering result of the image.
As a further limitation of the second aspect of the present invention, the obtaining, in the gray value sorting module, a pixel matrix centered on a certain pixel point, and sorting the one-dimensional sequences of the pixel matrix according to the gray value respectively to obtain the gray value median of each one-dimensional sequence includes:
configuring continuous serial numbers for all gray values of any one-dimensional sequence according to the sequence of entering a processing flow, and recording the result as 1 when the current gray value is larger than the gray value before the corresponding sequence of the gray value, or recording the result as 0; when the current gray value is greater than or equal to the gray value after the sequence corresponding to the gray value, recording the result as 1, otherwise, recording the result as 0;
comparing each gray value of the current one-dimensional sequence with other gray values of the current one-dimensional sequence in sequence to obtain a comparison and addition result corresponding to each gray value, and obtaining a gray value median of the current one-dimensional sequence by taking the comparison and addition result corresponding to each gray value as a sequencing serial number;
and performing the sorting on the gray values of the one-dimensional sequences to obtain the gray value median of the one-dimensional sequences.
As a further limitation of the second aspect of the present invention, the sorting module of gray-scale values sorts the median of gray-scale values of the one-dimensional sequences according to their magnitudes, and includes:
sorting all the gray value median values, and recording the result as 1 when the current gray value median value is larger than the gray value median value before the corresponding order of the gray value median value, or recording the result as 0; when the current gray value median is greater than or equal to the gray value median after the order corresponding to the gray value median, recording the result as 1, otherwise, recording the result as 0;
and comparing the gray value median with other gray value medias in sequence to obtain a comparison and addition result corresponding to each gray value median, taking the comparison and addition result corresponding to each gray value median as a sorting serial number to obtain the median of each gray value median, and taking the median of each gray value median as an updated value of the gray value of the central pixel point of the sliding window.
As a further limitation of the second aspect of the invention, the one-dimensional sequence is a row sequence or a column sequence.
A third aspect of the present invention provides a computer-readable storage medium, on which a program is stored, which, when being executed by a processor, carries out the steps of a method for median filtering of an image according to the first aspect of the present invention.
A fourth aspect of the present invention provides an electronic device, including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in the image median filtering method according to the first aspect of the present invention.
Compared with the prior art, the invention has the beneficial effects that:
1. the image median filtering method, the system, the medium and the electronic equipment adopt a parallel full-comparison sorting method, firstly carry out one-dimensional comparison and then carry out two-dimensional comparison, realize the quick search of the median of the gray values in the current window and greatly improve the speed of image median filtering processing.
2. The image median filtering method, the image median filtering system, the image median filtering medium and the electronic equipment fully adopt a parallel processing idea, utilize the advantages of pipeline processing and effectively solve the problem of poor real-time performance of the conventional median filtering scheme.
Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, are included to provide a further understanding of the invention, and are incorporated in and constitute a part of this specification, illustrate exemplary embodiments of the invention and together with the description serve to explain the invention and not to limit the invention.
Fig. 1 is a schematic diagram of an overall structure of two-dimensional sorting filtering provided in embodiment 1 of the present invention;
fig. 2 is a block diagram of one-dimensional sorting implementation provided in embodiment 1 of the present invention;
fig. 3 is a schematic diagram of a two-dimensional data sorting function provided in embodiment 1 of the present invention.
Detailed Description
The invention is further described with reference to the following figures and examples.
The embodiments and features of the embodiments of the invention may be combined with each other without conflict.
The parallel full-comparison sorting method adopts a parallel structure, and the basic principle is that the comparison results of all data and other data are completed at the same time, so that the time complexity is minimum, the sorting work can be completed only by one clock cycle, and the parallel full-comparison sorting method is very suitable for realizing FPGA (field programmable gate array) pipeline processing.
The invention first explains the principle of parallel comparison, assuming that n +1 data are to be sorted now, namely d0, d1, d2, · · ·, dn, then the processing steps for parallel sorting are as follows:
(1) The n +1 data are obtained simultaneously, and the n +1 data time sequence alignment is realized;
(2) And simultaneously comparing the n +1 data with other data respectively, recording comparison results, and defining: if the current data is larger than other data, recording the result as 1; if the current data is less than or equal to other data, recording the result as 0;
(3) Calculating the sum of all comparison results of each data in the step (2), wherein n comparison results are necessary as n +1 data are calculated;
(4) And (4) summing the comparison results in the step (3), namely the sorting result.
The following describes, in a more intuitive manner, the implementation steps of the algorithm, assuming that the following five data are to be sorted: 35. 19, 57, 41, 82;
suppose five data have already finished the time sequence alignment, namely can get these 5 data in the same cycle, next clock cycle, finish 5 data with other 4 data comparison at the same time separately, and record the result of comparison, namely the comparison value is 1 when the present data is greater than other data, otherwise is 0, the comparison result is as shown in table 1.
Table 1: comparison results
In the table, "x" indicates an irrelevant item, since it is meaningless compared with itself, and this value may be counted as 0. As can be seen from the above table, after comparison in one clock cycle, the order information of each data and other data is obtained, and the ordering information of each data in all data can be calculated by addition operation in the next clock cycle. As shown in the last row of the table. Due to the row-column consistency in the table above, only 2 × n comparators are needed to achieve the full comparison at this time.
It should be noted that if there are two data or more numbers in the middle that are equal, the following example is given for the case, such as a set of numbers as follows: 48 96, 51,6, 48;
the sorting was also performed according to the above sorting rules, as shown in table 2.
Table 2: comparison results
In the final row of the sorting result, the two data are identical in the summation and the summation is consistent, because the two data are "isotropic", so that it is obviously problematic in a pure counting manner, if the 5 data are identical, the last data is necessarily all 0, only one maximum value can be obtained, and the output of the median or other sorting cannot be obtained.
Therefore, for the same numerical value, it is necessary to find out "different characteristics" to be treated differently, and an obvious "different characteristic" is the input order of each numerical value, and is defined as follows:
(1) When the current data is larger than the input data before the current data, recording the result as 1, and recording the result as 0 when the result is smaller than or equal to the current data;
(2) When the current data is more than or equal to the input data after the current data, recording the result as 1, otherwise, recording the result as 0;
the data is reordered according to the above specification, as shown in table 3.
Table 3: results of the sorting
It can be seen that the data "48" input later has a low input priority, so the sorting result is reduced by one step, and the sorting purpose is just achieved.
Besides considering the influence of the same data on the ordering, attention needs to be paid to the problem of logic resource consumption after reordering, and inconsistency occurs in the ranks. Since greater than and greater than or equal to are different logics, taking the ordering of the 3 data d1, d2, d3 as an example, the comparison to be done is as follows: d1 is more than or equal to d2; d1 is more than or equal to d3; d2 > d1; d2 is more than or equal to d3; d3 > d1; d3 > d2.
Therefore, unless a separate equal sign judgment circuit is designed, each comparison is not overlapped, and then the number of needed comparators is n (n-1), 1 comparator occupies 5 logic units for calculation, the logic unit of the comparator occupies 5n (n-1), and in addition, n-1 comparison result registers and n-1 adders are needed for realizing the addition of the comparison results.
Example 1:
the embodiment provides a parallel full-comparison sequencing implementation scheme based on an FPGA implementation platform:
the parallel full-comparison sorting method adopts a design principle of changing speed by area, so that the sorting operation is very suitable for the pipeline operation, and since the sorting operation is of the same polarity in the row and column directions of the images, the sorting in the row direction of the one-dimensional images is considered in the same way, and then the sorting of the row sorting results in the column direction is considered, so that the sorting result in a window can be obtained, and similarly, the alignment in the row direction is realized by adopting row cache, as shown in fig. 1.
As can be seen from fig. 1, the design emphasis is on vector ordering in one-dimensional direction. Thus, operations in one dimension can be packed as separate modules. The line cache is responsible for aligning the sorting results of different lines, the design is mainly directed at the implementation of a one-dimensional sorting algorithm, the fully parallel sorting in the one-dimensional direction needs to simultaneously implement the mutual comparison of all data, and the results are added and then output, and the design of the one-dimensional sorting is shown in fig. 2.
The one-dimensional full parallel sequencing method comprises the following steps:
(1) Firstly, n data to be sequenced are obtained, which can be realized by beating n-1 beats in a data stream;
(2) And (3) carrying out full comparison: sequentially comparing the current data with all other data, and recording a comparison result, wherein the input sequence problem needs to be considered in the comparison process;
(3) Adding the recording results in (2): the addition operation can be completed by a plurality of clocks according to different comparison widths;
(4) And (4) searching the addition result in the step (3) to determine a sorting order, finding a middle sorting sequence number, extracting corresponding data and outputting the data, wherein the data is the median of the one-dimensional sequence.
The above steps for realizing one-dimensional data sorting are given, the median filtering processing is performed on the image, two-dimensional data sorting needs to be realized to obtain the pixel median of the current image window, the two-dimensional sorting is different from one-dimensional sorting, and more pipeline processing is needed, and the whole calculation steps are as follows:
(1) Calculating the sorting results of all one-dimensional row directions in the data window, and outputting the sorting results;
(2) Extracting a median value of each one-dimensional row sequence according to the sorting result to form a new one-dimensional sequence;
(3) Sequencing the new one-dimensional sequence, extracting sequence median data and outputting;
(4) Realizing time sequence alignment;
fig. 3 shows a two-dimensional data sorting function composition, where two-dimensional data is input line by line, sequentially written into read line buffers (0 to n-2), multiple line vectors are extracted and registered according to a set window size, one-dimensional line sorting processing is performed on each line vector, a median of each one-dimensional line vector is extracted, a new sequence is formed by the line vector medians, sorting processing is performed as a one-dimensional sequence, and a two-dimensional sequence median is obtained and output.
The method has the advantages that the image median filtering is realized by adopting a parallel full-comparison sorting method, the pixel value window of the image is regarded as a two-dimensional sequence, the two-dimensional sequence sorting is converted into a one-dimensional sequence sorting, the parallel processing advantages of the FPGA are fully utilized, the median filtering processing of the ultra-high definition image is realized by adopting a multi-stage pipeline mode, the method has good expansibility, and the processing real-time performance is greatly improved compared with the conventional method, for example, the output delay of a 3 multiplied by 3 window is 6 clock cycles, and the output delay of 5 multiplied by 5 is 9 clock cycles.
It may be understood that, in some other implementation manners, sequencing in the row direction may also be performed first (that is, each column sequence is a one-dimensional sequence), and then sequencing in the row direction may be performed, which may be selected by a person skilled in the art according to a specific working condition, and details are not described here.
Example 2:
an embodiment 2 of the present invention provides an image median filtering system, including:
a grayscale ranking module configured to: acquiring a pixel matrix taking a certain pixel point as a center, and sequencing each one-dimensional sequence of the pixel matrix according to the gray value to obtain the gray value median of each one-dimensional sequence; sorting the grey value median values of all the one-dimensional sequences according to the magnitude, and taking the median value of the grey value median sorting result as an update value of the grey value of the central pixel point of the pixel matrix;
a median filtering module configured to: and sequentially carrying out the processing on each pixel point of the image to obtain a median filtering result of the image.
The working method of the system is the same as the image median filtering method provided in embodiment 1, and details are not repeated here.
Example 3:
embodiment 3 of the present invention provides a computer-readable storage medium, on which a program is stored, which, when being executed by a processor, implements the steps in the image median filtering method according to embodiment 1 of the present invention.
Example 4:
embodiment 4 of the present invention provides an electronic device, which includes a memory, a processor, and a program stored in the memory and executable on the processor, where the processor executes the program to implement the steps in the image median filtering method according to embodiment 1 of the present invention.
Claims (10)
1. An image median filtering method, characterized by comprising the following processes:
acquiring a pixel matrix taking a certain pixel point as a center, and sequencing each one-dimensional sequence of the pixel matrix according to the gray value to obtain the gray value median of each one-dimensional sequence; sorting the gray value median values of all the one-dimensional sequences according to the size, and taking the median value of the gray value median sorting result as an update value of the gray value of a central pixel point of the pixel matrix;
and sequentially carrying out the processing on each pixel point of the image to obtain a median filtering result of the image.
2. The method of claim 1,
acquiring a pixel matrix with a certain pixel point as a center, wherein each one-dimensional sequence of the pixel matrix is sorted according to the gray value, and the gray value median of each one-dimensional sequence is obtained, and the method comprises the following steps:
configuring continuous serial numbers for all gray values of any one-dimensional sequence according to the sequence of entering a processing flow, and recording the result as 1 when the current gray value is larger than the gray value before the corresponding sequence of the gray value, or recording the result as 0; when the current gray value is greater than or equal to the gray value after the sequence corresponding to the gray value, recording the result as 1, otherwise, recording the result as 0;
comparing each gray value of the current one-dimensional sequence with other gray values of the current one-dimensional sequence in sequence to obtain a comparison and addition result corresponding to each gray value, and obtaining a gray value median of the current one-dimensional sequence by taking the comparison and addition result corresponding to each gray value as a sorting sequence number;
and performing the sorting on the gray values of the one-dimensional sequences to obtain the gray value median of the one-dimensional sequences.
3. The method of claim 1,
sorting the gray value median values of the one-dimensional sequences according to size, and comprising the following steps:
sorting all the gray value median values, and recording the result as 1 when the current gray value median value is larger than the gray value median value before the corresponding order of the gray value median value, or recording the result as 0; when the current gray value median is greater than or equal to the gray value median after the order corresponding to the gray value median, recording the result as 1, otherwise, recording the result as 0;
and comparing the median of each gray value with the median of other gray values in sequence to obtain a comparison and addition result corresponding to the median of each gray value, taking the comparison and addition result corresponding to the median of each gray value as a sorting serial number to obtain the median of each gray value, and taking the median of each gray value as an update value of the gray of the central pixel point of the sliding window.
4. The method of any one of claims 1-3, wherein,
the one-dimensional sequence is a row sequence or a column sequence.
5. An image median filtering system, comprising:
a grayscale value ordering module configured to: acquiring a pixel matrix taking a certain pixel point as a center, and sequencing each one-dimensional sequence of the pixel matrix according to the gray value to obtain the gray value median of each one-dimensional sequence; sorting the gray value median values of all the one-dimensional sequences according to the size, and taking the median value of the gray value median sorting result as an update value of the gray value of a central pixel point of the pixel matrix;
a median filtering module configured to: and sequentially carrying out the processing on each pixel point of the image to obtain a median filtering result of the image.
6. The image median filtering system of claim 5,
in the gray value sorting module, a pixel matrix with a certain pixel point as a center is obtained, and each one-dimensional sequence of the pixel matrix is sorted according to the gray value size to obtain the gray value median of each one-dimensional sequence, including:
configuring continuous serial numbers for all gray values of any one-dimensional sequence according to the sequence of entering a processing flow, and recording the result as 1 when the current gray value is larger than the gray value before the corresponding sequence of the gray value, or recording the result as 0; when the current gray value is greater than or equal to the gray value after the sequence corresponding to the gray value, recording the result as 1, otherwise, recording the result as 0;
comparing each gray value of the current one-dimensional sequence with other gray values of the current one-dimensional sequence in sequence to obtain a comparison and addition result corresponding to each gray value, and obtaining a gray value median of the current one-dimensional sequence by taking the comparison and addition result corresponding to each gray value as a sequencing serial number;
and performing the sorting on the gray values of the one-dimensional sequences to obtain the gray value median of the one-dimensional sequences.
7. The image median filtering system of claim 5,
in the gray value sorting module, sorting the gray value median values of the one-dimensional sequences according to size comprises:
sorting all the gray value median values, and recording the result as 1 when the current gray value median value is larger than the gray value median value before the corresponding order of the gray value median value, or recording the result as 0; when the current gray value median is greater than or equal to the gray value median after the order corresponding to the gray value median, recording the result as 1, otherwise, recording the result as 0;
and comparing the median of each gray value with the median of other gray values in sequence to obtain a comparison and addition result corresponding to the median of each gray value, taking the comparison and addition result corresponding to the median of each gray value as a sorting serial number to obtain the median of each gray value, and taking the median of each gray value as an update value of the gray of the central pixel point of the sliding window.
8. The image median filtering system of any one of claims 5 to 7,
the one-dimensional sequence is a row sequence or a column sequence.
9. A computer-readable storage medium, on which a program is stored, which, when being executed by a processor, carries out the steps of a method for median filtering of an image as claimed in any one of claims 1 to 4.
10. An electronic device comprising a memory, a processor and a program stored in the memory and executable on the processor, wherein the processor implements the steps of a method for median filtering of images as claimed in any one of claims 1 to 4 when executing the program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310238322.9A CN115953335A (en) | 2023-03-14 | 2023-03-14 | Image median filtering method, system, medium and electronic device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310238322.9A CN115953335A (en) | 2023-03-14 | 2023-03-14 | Image median filtering method, system, medium and electronic device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115953335A true CN115953335A (en) | 2023-04-11 |
Family
ID=87282805
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310238322.9A Pending CN115953335A (en) | 2023-03-14 | 2023-03-14 | Image median filtering method, system, medium and electronic device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115953335A (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109785267A (en) * | 2019-01-21 | 2019-05-21 | 哈尔滨理工大学 | A kind of improved Fast Median Filtering algorithm |
| CN110738106A (en) * | 2019-09-05 | 2020-01-31 | 天津大学 | optical remote sensing image ship detection method based on FPGA |
-
2023
- 2023-03-14 CN CN202310238322.9A patent/CN115953335A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109785267A (en) * | 2019-01-21 | 2019-05-21 | 哈尔滨理工大学 | A kind of improved Fast Median Filtering algorithm |
| CN110738106A (en) * | 2019-09-05 | 2020-01-31 | 天津大学 | optical remote sensing image ship detection method based on FPGA |
Non-Patent Citations (3)
| Title |
|---|
| 师廷伟 金长江: "基于FPGA的并行全比较排序算法", 《数字技术与应用》 * |
| 林琦琛: "基于FPGA的PET图像快速重建与预处理系统研究", 《中国优秀硕士学位论文全文数据库》 * |
| 苟旭辉: "基于FPGA的带式输送机光学动态流量计", 《中国优秀硕士学位论文全文数据库 基础科技辑》 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112613581B (en) | Image recognition method, system, computer equipment and storage medium | |
| Vega-Rodríguez et al. | An FPGA-based implementation for median filter meeting the real-time requirements of automated visual inspection systems | |
| CN109784372B (en) | Target classification method based on convolutional neural network | |
| He et al. | FPGA-based real-time super-resolution system for ultra high definition videos | |
| CN106682258B (en) | Multi-operand addition optimization method and system in high-level comprehensive tool | |
| CN113362277A (en) | Workpiece surface defect detection and segmentation method based on deep learning | |
| Kalali et al. | Low complexity 2D adaptive image processing algorithm and its hardware implementation | |
| CN113313252B (en) | Depth separable convolution implementation method based on pulse array | |
| JP4310500B2 (en) | Important component priority calculation method and equipment | |
| CN115953335A (en) | Image median filtering method, system, medium and electronic device | |
| Meher et al. | Hardware-efficient systolic-like modular design for two-dimensional discrete wavelet transform | |
| Yu et al. | A memory-efficient hardware architecture for deformable convolutional networks | |
| Pang et al. | Design and implement of median filter toward remote sensing images based on FPGA | |
| CN112508174B (en) | Weight binary neural network-oriented pre-calculation column-by-column convolution calculation unit | |
| CN114519425A (en) | Convolution neural network acceleration system with expandable scale | |
| CN110765413B (en) | Matrix summation structure and neural network computing platform | |
| Bevara et al. | VLSI implementation of high throughput parallel pipeline median finder for IoT applications | |
| Matsubara et al. | A low-complexity and low power median filter design | |
| CN112001492A (en) | Mixed flow type acceleration framework and acceleration method for binary weight Densenet model | |
| CN111079904A (en) | Acceleration method of deep separable convolution, storage medium and application | |
| Zhang et al. | Multi-level Feature Fusion Network for Single Image Super-Resolution | |
| Kayala et al. | Pipelined Shapiro sorting for image applications | |
| Tasdizen et al. | Computation reduction techniques for vector median filtering and their hardware implementation | |
| Hsiao et al. | A novel FPGA architectural implementation of pipelined thinning algorithm | |
| Hernandez et al. | A combined VLSI architecture for nonlinear image processing filters |
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: 20230411 |