CN110163868B - Remote sensing image segmentation method - Google Patents
Remote sensing image segmentation method Download PDFInfo
- Publication number
- CN110163868B CN110163868B CN201910297705.7A CN201910297705A CN110163868B CN 110163868 B CN110163868 B CN 110163868B CN 201910297705 A CN201910297705 A CN 201910297705A CN 110163868 B CN110163868 B CN 110163868B
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- features
- sensing image
- segmentation
- 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.)
- Active
Links
- 238000003709 image segmentation Methods 0.000 title claims abstract description 36
- 238000000034 method Methods 0.000 title claims abstract description 36
- 230000011218 segmentation Effects 0.000 claims abstract description 30
- 238000001914 filtration Methods 0.000 claims abstract description 10
- 230000004927 fusion Effects 0.000 claims description 5
- 238000007781 pre-processing Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 abstract description 7
- 238000007500 overflow downdraw method Methods 0.000 abstract description 2
- 238000000605 extraction Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—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
- G06T2207/10032—Satellite or aerial image; Remote sensing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a remote sensing image segmentation method, which comprises the following steps: carrying out noise reduction pretreatment on the collected remote sensing image to obtain a denoised remote sensing image; carrying out gray level processing and super-pixel segmentation on the remote sensing image after the noise is removed; respectively extracting texture features and color features based on the superpixel segmentation result; extracting geometric features of the remote sensing image subjected to gray level processing; fusing the texture features, the color features and the geometric features to obtain image segmentation features; based on image segmentation characteristics, performing image filtering by adopting mean shift to obtain a primary segmentation result; and carrying out region combination on the initial segmentation result to obtain a final segmentation result, fully considering the processing of noise, and adopting a multi-feature fusion method to segment the image, so that the image segmentation effect can be fully ensured.
Description
Technical Field
The invention relates to the technical field of remote sensing image processing, in particular to a remote sensing image segmentation method.
Background
Image segmentation refers to the technique and process of dividing an image into regions with features and extracting an object of interest. Because it is the key step of image analysis, computer vision and pattern recognition, so it is always highly valued by people, and thousands of segmentation algorithms have been proposed so far, but most of them are based on the segmentation of gray level image, because the color image provides more abundant information than gray level image, so as to increase the processing ability of computer, people pay more and more attention to the segmentation of color image. The remote sensing image is a very common and typical color image, is widely concerned, and is a research hotspot of broad scholars.
Noise generated in the process of acquiring and transmitting a remote sensing image is usually not one, in the existing remote sensing image segmentation method, the research is mainly focused on how to realize more accurate segmentation, a pre-processing part at the early stage is omitted, the image pre-processing result is not good, the subsequent image segmentation effect is directly influenced, in addition, the color and space information of the remote sensing image is only considered in the current segmentation of the large-square part image, the segmentation effect on the image with rich texture is not ideal, and the segmentation effect cannot be ensured.
Therefore, how to sufficiently ensure the image segmentation effect is a problem that needs to be solved urgently by those skilled in the art.
Disclosure of Invention
In view of this, the invention provides a remote sensing image segmentation method, which takes processing of noise into full consideration, and adopts a multi-feature fusion method to segment an image, so as to fully ensure the image segmentation effect.
In order to achieve the purpose, the invention adopts the following technical scheme:
a remote sensing image segmentation method comprises the following steps:
carrying out noise reduction pretreatment on the acquired remote sensing image to obtain a denoised remote sensing image;
carrying out gray processing and super-pixel segmentation on the remote sensing image subjected to noise removal;
respectively extracting texture features and color features based on the super-pixel segmentation result; the method specifically comprises the following steps: based on a super-pixel segmentation result, removing textures by using a filtering method based on a total variation model to obtain color characteristics, and filtering by using a Gabor filter to obtain texture characteristics;
extracting geometric features of the remote sensing image subjected to gray level processing;
fusing the texture features, the color features and the geometric features to obtain image segmentation features;
based on image segmentation characteristics, performing image filtering by adopting mean shift to obtain a primary segmentation result;
and carrying out region combination on the preliminary segmentation result to obtain a final segmentation result.
Preferably, the method comprises the following steps: the noise reduction pretreatment comprises the following specific steps:
s1: judging whether the absolute value of the difference between the gray value of the pixel point and the line mean value is greater than a threshold value, if so, determining that the pixel point is impulse noise, eliminating the point, and outputting the line mean value as the pixel value of the point; otherwise, the pixel value of the point is not changed, and a first denoising result is obtained;
s2: judging whether the pixel points are edge pixel points or not based on the first denoising result, the formula (1) and a preset threshold value;
g(i,j)=|f(i,j)+f(i,j+1)|-|f(i,j-1)+f(i,j-2)| (1)
wherein f (i, j) is a selected pixel point, f (i, j + 1), f (i, j-1) and f (i, j-2) are adjacent pixel points of the selected pixel point, g (i, j) is a gradient value, when g (i, j) is greater than a preset threshold value, the selected point is judged to be an edge point, and the edge point is directly output; otherwise, judging that the selected point is not an edge point, and outputting a line mean value.
Preferably, the selection of the preset threshold value T is determined according to the brightness characteristics of the image, and T = -0.5 × MV + a;
wherein MV is a row mean value and a is a constant.
Preferably, a =100.
Preferably, the texture feature, the color feature and the geometric feature are fused by using a linear fusion technology.
Preferably, the texture feature, the color feature and the geometric feature are subjected to Bayesian fusion.
Preferably, the color characteristics are specifically: lab color characteristics.
According to the technical scheme, compared with the prior art, the remote sensing image segmentation method provided by the invention has the advantages that the processing of noise is fully considered, and the interference of the noise on the subsequent feature extraction is prevented, so that the inaccuracy of the feature extraction is avoided. In addition, in the remote sensing image segmentation method provided by the invention, the color feature, the texture feature and the geometric feature of the remote sensing image are respectively extracted, and various features are fused, so that the remote sensing image segmentation result is more accurate and more guaranteed.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
FIG. 1 is a flow chart of a remote sensing image segmentation method provided by the invention;
FIG. 2 is a flow chart of denoising provided by the present invention;
FIG. 3 is a diagram illustrating the experimental effects of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
With reference to the accompanying drawing 1, the embodiment of the invention discloses a remote sensing image segmentation method, which specifically comprises the following steps:
s1: carrying out noise reduction pretreatment on the acquired remote sensing image to obtain a denoised remote sensing image;
s2: carrying out gray level processing and super-pixel segmentation on the remote sensing image after the noise is removed;
s3: respectively extracting texture features and color features based on the super-pixel segmentation result; the method specifically comprises the following steps: based on a super-pixel segmentation result, removing textures by using a filtering method based on a total variation model to obtain color features, and filtering by using a Gabor filter to obtain texture features;
s4: extracting geometric features of the remote sensing image subjected to gray level processing;
s5: fusing the texture features, the color features and the geometric features to obtain image segmentation features;
s6: based on image segmentation characteristics, performing image filtering by adopting mean shift to obtain a primary segmentation result;
s7: and carrying out region combination on the preliminary segmentation result to obtain a final segmentation result.
In the remote sensing image segmentation method provided by the invention, the processing of noise is fully considered, and the inaccuracy of feature extraction caused by the interference of noise on the subsequent feature extraction is prevented. In addition, in the remote sensing image segmentation method provided by the invention, the color feature, the texture feature and the geometric feature of the remote sensing image are respectively extracted, and various features are fused, so that the remote sensing image segmentation result is more accurate and more guaranteed.
In order to further optimize the technical scheme, the method comprises the following steps: the noise reduction pretreatment comprises the following specific steps:
s11: judging whether the absolute value of the difference between the gray value of the pixel point and the line mean value is greater than a threshold value, if so, determining that the pixel point is impulse noise, excluding the point, and outputting the line mean value as the pixel value of the point; otherwise, the pixel value of the point is not changed, and a first denoising result is obtained;
s12: judging whether the pixel points are edge pixel points or not based on the first denoising result, the formula (1) and a preset threshold value;
g(i,j)=|f(i,j)+f(i,j+1)|-|f(i,j-1)+f(i,j-2)| (1)
wherein f (i, j) is the selected pixel point, f (i, j + 1), f (i, j-1) and f (i, j-2) are adjacent pixel points of the selected pixel point, g (i, j) is a gradient value, when g (i, j) is larger than a preset threshold value, the selected point is judged to be an edge point, and the edge point is directly output; otherwise, judging that the selected point is not the edge point, and outputting the line average value.
In order to further optimize the technical scheme, the selection of the preset threshold value T is determined according to the brightness characteristics of the image, and T = -0.5 multiplied by MV + a;
wherein MV is a row mean value and a is a constant.
In order to further optimize the above solution, a =100.
a is a constant and can be set according to specific experimental conditions.
In order to further optimize the technical scheme, a linear fusion technology is adopted to fuse the texture feature, the color feature and the geometric feature.
In order to further optimize the technical scheme, bayesian fusion is carried out on the texture features, the color features and the geometric features.
In order to further optimize the technical scheme, the color characteristics are specifically as follows: lab color characteristics.
In addition, the embodiment of the invention further elaborates the technical scheme of the invention by combining the specific experimental results.
With reference to fig. 3, it can be seen from fig. 3 that the method provided by the present invention can well segment different objects such as rivers, lands, civil houses, roads, etc., and further proves the effectiveness of the remote sensing image segmentation method provided by the present invention.
In the present specification, the embodiments are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed in the embodiment corresponds to the method disclosed in the embodiment, so that the description is simple, and the relevant points can be referred to the description of the method part.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (7)
1. A remote sensing image segmentation method is characterized by comprising the following steps:
carrying out noise reduction pretreatment on the collected remote sensing image to obtain a denoised remote sensing image;
carrying out gray level processing and super-pixel segmentation on the remote sensing image after the noise is removed;
respectively extracting texture features and color features based on the super-pixel segmentation result; the method specifically comprises the following steps: based on a super-pixel segmentation result, removing textures by using a filtering method based on a total variation model to obtain color features, and filtering by using a Gabor filter to obtain texture features;
extracting geometric features of the remote sensing image subjected to gray level processing;
fusing the texture features, the color features and the geometric features to obtain image segmentation features;
based on image segmentation characteristics, performing image filtering by adopting mean shift to obtain a primary segmentation result;
and carrying out region combination on the preliminary segmentation result to obtain a final segmentation result.
2. The remote sensing image segmentation method according to claim 1, wherein the noise reduction preprocessing comprises the following specific steps:
s1: judging whether the absolute value of the difference between the gray value of the pixel point and the line mean value is greater than a threshold value, if so, determining that the pixel point is impulse noise, excluding the point, and outputting the line mean value as the pixel value of the point; otherwise, the pixel value of the point is not changed, and a first denoising result is obtained;
s2: judging whether the pixel points are edge pixel points or not based on the first denoising result, the formula (1) and a preset threshold value;
g(i,j)=|f(i,j)+f(i,j+1)|-|f(i,j-1)+f(i,j-2)| (1)
wherein f (i, j) is the selected pixel point, f (i, j + 1), f (i, j-1) and f (i, j-2) are adjacent pixel points of the selected pixel point, g (i, j) is a gradient value, when g (i, j) is larger than a preset threshold value, the selected point is judged to be an edge point, and the edge point is directly output; otherwise, judging that the selected point is not the edge point, and outputting the line average value.
3. The remote sensing image segmentation method according to claim 2, wherein the selection of the preset threshold value T is determined according to the brightness characteristics of the image, T = -0.5 x MV + a;
wherein MV is a row mean value and a is a constant.
4. A remote sensing image segmentation method as claimed in claim 3, characterized in that a =100.
5. The remote sensing image segmentation method according to claim 1, wherein a linear fusion technique is used to fuse the texture features, the color features and the geometric features.
6. The remote sensing image segmentation method according to claim 1, wherein the texture features, the color features, and the geometric features are bayesian fused.
7. The remote sensing image segmentation method according to claim 1, wherein the color features are specifically: lab color characteristics.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910297705.7A CN110163868B (en) | 2019-04-17 | 2019-04-17 | Remote sensing image segmentation method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910297705.7A CN110163868B (en) | 2019-04-17 | 2019-04-17 | Remote sensing image segmentation method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110163868A CN110163868A (en) | 2019-08-23 |
CN110163868B true CN110163868B (en) | 2023-03-28 |
Family
ID=67639381
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910297705.7A Active CN110163868B (en) | 2019-04-17 | 2019-04-17 | Remote sensing image segmentation method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110163868B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112541920A (en) * | 2019-09-23 | 2021-03-23 | 大连民族大学 | Image super-pixel target pedestrian segmentation method based on multi-channel mode |
CN117095299B (en) * | 2023-10-18 | 2024-01-26 | 浙江省测绘科学技术研究院 | Grain crop extraction method, system, equipment and medium for crushing cultivation area |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105913421B (en) * | 2016-04-07 | 2018-11-16 | 西安电子科技大学 | Remote sensing images cloud detection method of optic based on adaptive shape dark |
CN107085833B (en) * | 2017-04-13 | 2019-07-16 | 长安大学 | Remote sensing images filtering method based on the equal intermediate value fusion of gradient inverse self-adaptive switch |
CN107578035B (en) * | 2017-09-30 | 2020-06-16 | 深圳市颐通科技有限公司 | Human body contour extraction method based on super-pixel-multi-color space |
-
2019
- 2019-04-17 CN CN201910297705.7A patent/CN110163868B/en active Active
Also Published As
Publication number | Publication date |
---|---|
CN110163868A (en) | 2019-08-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103020965B (en) | A kind of foreground segmentation method based on significance detection | |
CN110163219B (en) | Target detection method based on image edge recognition | |
CN106682665B (en) | Seven-segment type digital display instrument number identification method based on computer vision | |
TW200834459A (en) | Video object segmentation method applied for rainy situations | |
CN110163868B (en) | Remote sensing image segmentation method | |
CN105678723A (en) | Multi-focus image fusion method based on sparse decomposition and differential image | |
CN105719275A (en) | Parallel combination image defect segmentation method | |
CN109949294A (en) | A kind of fracture apperance figure crack defect extracting method based on OpenCV | |
CN105931259A (en) | High voltage transmission line extraction method based on morphology processing and device | |
CN110782385A (en) | Image watermark removing method based on deep learning | |
CN115731257A (en) | Leaf form information extraction method based on image | |
CN109087347B (en) | Image processing method and device | |
KR101615479B1 (en) | Method and apparatus for processing super resolution image using adaptive pre/post-filtering | |
CN113628129A (en) | Method for removing shadow of single image by edge attention based on semi-supervised learning | |
Nandy et al. | A comparative analysis of application of Niblack and Sauvola binarization to retinal vessel segmentation | |
CN109426770B (en) | Iris identification method | |
Kezia et al. | A color-texture based segmentation method to extract object from background | |
CN116205939A (en) | Line extraction method, line extraction apparatus, and computer storage medium | |
JP6044138B2 (en) | Image region dividing apparatus, method, and program | |
CN106803053B (en) | Fingerprint image processing method and device | |
CN114519694A (en) | Seven-segment digital tube liquid crystal display screen identification method and system based on deep learning | |
CN110430340B (en) | Method and system for reducing noise of pulse array signal | |
Erofeev et al. | Automatic logo removal for semitransparent and animated logos | |
Verma et al. | A fuzzy impulse noise filter based on boundary discriminative noise detection | |
CN110675372A (en) | Brain area automatic segmentation method and system of brain tissue three-dimensional image with horizontal cell resolution |
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 |