CN101710416A - Processing method for multiple-target remote sensing image clouds - Google Patents
Processing method for multiple-target remote sensing image clouds Download PDFInfo
- Publication number
- CN101710416A CN101710416A CN200910113560A CN200910113560A CN101710416A CN 101710416 A CN101710416 A CN 101710416A CN 200910113560 A CN200910113560 A CN 200910113560A CN 200910113560 A CN200910113560 A CN 200910113560A CN 101710416 A CN101710416 A CN 101710416A
- Authority
- CN
- China
- Prior art keywords
- image
- cloud
- images
- cloudless
- cloudy
- 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
Images
Landscapes
- Image Processing (AREA)
Abstract
The invention relates to a processing method for multiple-target remote sensing image clouds, comprising the following steps of: strictly registering two cloudy and unclouded images in the approximate order; carrying out histogram match on the registered images; masking the cloudy image; extracting an image of a cloudy region in the cloudy image corresponding to the unclouded image and extracting an image of an unclouded part in the cloudy image; carrying out compound operation on the images; de-noising smoothly and generating a result plot. The method realizes approximately same gray level distribution of the two images through the histogram match, thereby not only removing cloud covers from the images, but also keeping the original image information intact; some dead pixels left in the images can be easily processed by a smooth algorithm, thereby a clearer image is obtained; the image data are preprocessed simply and effectively, and the step operation is simple and clear.
Description
Technical field
The present invention relates to a kind of disposal route of multiple-target remote sensing image clouds, relate in particular to remote sensing image and remove cloud method.
Background technology
Develop rapidly along with infotech and sensor technology, modern remotely-sensed data is very abundant, satellite remote-sensing image resolution just develops towards higher spatial resolution, spectral resolution and temporal resolution direction, and remote sensing image has been widely used in a plurality of fields such as agriculture evaluation, geologic examination, environmental monitoring.But because weather, obtain fully that cloudless remote sensing image is the comparison difficulty sometimes, the shade of most of image when obtaining Shi Douhui being subjected to cloud and cloud and throwing more or less and the influence of gasoloid etc. on ground.This has brought trouble for the application person of many remote sensing images, the influence of how to remove cloud from remote sensing image, the matter of utmost importance that often many application persons faced.Common remote sensing images go cloud method to mainly contain: polynomial expression corrects method, digital elevation model and proofreaies and correct the image co-registration method of going cloud method, wavelet decomposition, homomorphic filtering method etc.
Polynomial expression corrects and to go the cloud method: be utilize polynomial expression with another width of cloth image rectification to benchmark image, to reach the purpose of cloud.The main thought of this method is: a cloudless image choosing identical area then, with two width of cloth image polynomial expression registrations, is replaced cloud-capped topography with cloudless image as reference map, reaches the purpose of target area image being removed cloud.The shortcoming one of this method be coefficient determine to exist certain difficulty; The 2nd, the quantity of registration atural object reference point depends on multinomial coefficient, if the cloud of target area distributes more for a long time, it is just bigger that polynomial expression corrects the image registration workload of removing cloud, and piece image has a plurality of polynomial supports possibly.
Digital elevation model is proofreaied and correct and is gone the cloud method: be the lifting effect at the higher regional landform of height above sea level, the method of the removal cloud that proposes at the overhead cloud body that forms of massif, this method is to use the figure with identical Geographic Reference to remove to proofread and correct the remote sensing images that cloud is arranged, and reaches the purpose of removing cloud.
Homomorphic filtering method: be to use illumination/Reflectivity Model remote sensing images are carried out Filtering Processing, usually be applied to disclose the minutia of shadow region, but on algorithm, realize more complicated that processing speed is slow.
Concerning the image that on a large scale, has thin cloud, adopt the homomorphic filtering method better.For thicker cloud layer, cannot based on the cloud processing of going of single scape image.Therefore, can adopt the complementary information of the image of multidate to go cloud to handle, that is: utilize the cloudless image of close areal of time that the cloud sector is substituted, thereby reach the effect of cloud.
Summary of the invention
The object of the invention is, a kind of disposal route of multiple-target remote sensing image clouds is provided, this method with two width of cloth times close cloud and the cloudless strict registration of image arranged; Image behind the registration is carried out the histogram coupling; The image that cloud is arranged is carried out mask; Extracting in the cloudless image should have the image that the territory, cloud sector is arranged in the cloud atlas picture relatively, extracts the image that cloudless part in the cloud atlas picture is arranged; The image compound operation; Smoothing denoising generates figure as a result.It is roughly the same that this method realizes that by the histogram coupling two gradation of images distribute, thereby reach the cloud layer that can reject in the image, can guarantee the seamless spliced cloud effect of going that original image information is intact again; At some bad points of leaving in the image, be easy to just handle with level and smooth algorithm, thereby obtain more distinct image; Pre-processing image data is simply effective, and step is simple to operate to be understood.
The disposal route of multiple-target remote sensing image clouds of the present invention follows these steps to carry out:
A, with two width of cloth times close cloud and the cloudless strict registration of image arranged;
B, the image behind the registration is carried out histogram coupling, the histogram distribution of coupling back two width of cloth images is similar;
C, the image that cloud is arranged is carried out mask, generate two width of cloth bianry images (A) and (B), have the regional value white portion of cloud to be (1) in image (A), cloud-free area thresholding black region is (0); Have the regional value black region of cloud to be (0) in the image (B), cloud-free area thresholding white portion is (1);
Relatively the image that the territory, cloud sector is arranged in the cloud atlas picture should be arranged in d, the cloudless image of extraction, extraction has the image of cloudless part in the cloud atlas picture, carry out computing with image (A) and cloudless image, extract the image (C) that the territory, cloud sector is arranged in the cloud atlas picture should be arranged in the cloudless image relatively, with image (B) with have cloud atlas to look like to carry out computing, extract the image (D) that cloudless part in the cloud atlas picture is arranged;
E, image compound operation, with image (C) with (D) superpose computing, form a complete image (E) that goes behind the cloud;
Total some noise spot of the image that forms among f, the step e is removed noise spot by selecting smoothing filter, the image of step e is gone bad a little repaired again, gets the last F of figure as a result.
The disposal route of multiple-target remote sensing image clouds of the present invention because state, posture and solar irradiation, atmospheric diffusion and the absorption of sensor all can cause the radiation difference of atural object, has tangible boundary line between the image of filling and the map.Therefore how eliminating the boundary line is a main problem.For realizing the non-boundary between image, two kinds of methods are arranged generally.The one, directly in existing software, directly two width of cloth images are carried out the histogram coupling; The 2nd, pseudo-invariant features method.Realize that pseudo-invariant features method generally needs 3 steps:
At first, select radiation value in two width of cloth images not have or the target area (being called invariant region) of rare variation;
Use the average radiation value of invariant region then, find the solution in the linear equality (1) of radiation value between image parameter and, obtain the linear relationship of radiation value;
At last,,, finish relative radiant correction by linear transformation to radiation value according to the linear relationship that obtains between radiation value, consistent on the radiation value that makes image to be corrected and the reference picture.Therefore, for realizing the relative radiant correction of remote sensing images, just must solve 3 problems: the 1. selection of invariant region; 2. linear relation finds the solution; 3. the conversion of gradation of image.
These two kinds of methods all are based on principle of uniformity.Can make the Luminance Distribution of two width of cloth images approaching as much as possible by the histogram coupling, can eliminate radiation difference like this, be that the image after merging does not have tangible boundary line.The histogram matching principle is as follows:
By the histogram of source images and the histogram of target image are mated, make itself and target image that consistent histogram distribution be arranged, reach the purpose in trimming boundary line.
Suppose not that the gradation of image value of phase is asked simultaneously and satisfy linear relationship, this hypothesis is set up under approximate situation.At this moment can gray-scale relation in the phase images simultaneously be described not by linear equality.In processing procedure, each wave band has been generated a slit mask file, data available in the image is labeled as 1, and the data markers that will need to fill is 0.In case determined to have the position in territory, cloud sector, just can utilize linear histogram matching method, between two width of cloth images, set up a linear transformation, correcting gain that conversion is used and biasing can calculate by the average and the standard variance of view data.
For certain any pixel value x in the cloudless image with the pixel value Y of same point in the cloud atlas picture is arranged, the linear transformation of foundation can equal GX with formula Y and add B, and wherein G is the correcting gain that is used for the histogram coupling, and B is the biasing that is used for the histogram coupling.This conversion is applied in the view picture blank map picture, is the histogram matching method.In the histogram matching process, G and B are the values of fixing, thereby therefore the purpose that has each picture element that need fill of cloud part to reach cloud in the cloud atlas picture is arranged.
Description of drawings
Fig. 1 is a process flow diagram of the present invention
Fig. 2 for two width of cloth times of the present invention close cloud and cloudless striograph arranged, wherein 1 for there being cloud atlas, 2 is no cloud atlas
Fig. 3 is cloud mask figure of the present invention (A), background mask figure (B)
Fig. 4 extracts the figure (C) of cloud sector correspondence for the present invention
The Background (D) that Fig. 5 extracts for the present invention
Fig. 6 is the figure (E) of the present invention after compound
Fig. 7 is the figure as a result (F) behind the present invention smoothly
Embodiment
Embodiment
Be the purpose that realizes removing cloud, the present invention mainly utilizes the IDL programming software to realize, further describes with two width of cloth striographs
A, with two width of cloth times close cloud and the cloudless strict registration of image arranged;
B, the image behind the registration is carried out histogram coupling, the histogram distribution of coupling back two width of cloth images is similar;
C, the image that cloud is arranged is carried out mask, generate two width of cloth bianry image A and B, the regional value white portion that cloud is arranged in image A is 1, and cloud-free area thresholding black region is 0; The regional value black region that cloud is arranged in the image B is 0, cloud-free area thresholding white portion is 1, and in envi software, setting up a mask extraction has cloud areal map A, set up an anti-mask extraction figure B again, realize by setting up a simple expression formula (b1 eq 0) * 1or (b1 eq 0) * 0.When the value of b1 is 0 it is changed to 1, its value is to change 0 at 1 o'clock, realizes confuse right and wrong.Especially note when setting up mask, suitably select threshold value to extract territory, cloud sector (threshold value in cloud sector is made as 60-255 in the present invention's experiment);
Relatively the image that the territory, cloud sector is arranged in the cloud atlas picture should be arranged in d, the cloudless image of extraction, extraction has the image of cloudless part in the cloud atlas picture, carry out computing with image A and cloudless image, extract in the cloudless image and relatively the image C that the territory, cloud sector is arranged in the cloud atlas picture should be arranged, go and have cloud atlas to look like to carry out computing with image B, extraction has the image D of cloudless part in the cloud atlas picture, and directly using mask can obtain corresponding image in envi software;
E, image compound operation, image C and D are carried out computing be superimposed upon one, form a complete image E who goes behind the cloud, undertaken compound by a wave band computing this two width of cloth image C, D, expression formula is (b1 eq 0) * b2 or (b1 ne 0) * b1, when the value of picture dot among the wave band b1 is 0, be that 0 picture dot replaces to the pairing pixel value of wave band b2 with value; When the value of picture dot among the wave band b1 is not 0, do not replace, wherein the b1 correspondence is image C, the b2 correspondence be image D;
Total some noise spot of the image that forms among f, the step e, remove noise spot by selecting smoothing filter, again the image of step e is gone bad a little and repaired, get the last F of figure as a result, in envi software, can finish expression formula: (b1gt 200) * smooth (b1,3) or (b1 le 200) * b1 with an expression formula, if this pixel value greater than 200, is then filled this value with the average of 8 adjacent pixel values of its periphery; If value so just can be eliminated noise spot less than 200 no changes, reach the effect that needs.
See from the figure as a result that handles that the zone that cloud is arranged among the former figure is well filled up and gone up corresponding atural object, and eliminated the boundary line, reach the visual effect that needs.
The disposal route of multiple-target remote sensing image clouds of the present invention mainly is to utilize not the complementary characteristic of phase images information simultaneously to reach the best visual effect to a certain regional aim.The information that can keep former figure to greatest extent recovers to have the information in territory, cloud sector as far as possible when not losing former figure information.It is roughly the same to realize that by the histogram coupling two gradation of images distribute, thereby reaches the cloud layer that can reject in the image, can guarantee the seamless spliced cloud effect of going that original image information is intact again; This method is not limited to not the image of phase simultaneously, and the multispectral image data that it obtains the same area different sensors realize going cloud to handle, and this method can both realize in existing various image processing softwares.
Claims (1)
1. the disposal route of a multiple-target remote sensing image clouds is characterized in that following these steps to carrying out:
A, with two width of cloth times close cloud and the cloudless strict registration of image arranged;
B, the image behind the registration is carried out histogram coupling, the histogram distribution of coupling back two width of cloth images is similar;
C, the image that cloud is arranged is carried out mask, generate two width of cloth bianry images (A) and (B), have the regional value white portion of cloud to be (1) in image (A), cloud-free area thresholding black region is (0); Have the regional value black region of cloud to be (0) in the image B, cloud-free area thresholding white portion is (1);
Relatively the image that the territory, cloud sector is arranged in the cloud atlas picture should be arranged in d, the cloudless image of extraction, extraction has the image of cloudless part in the cloud atlas picture, carry out computing with image (A) and cloudless image, extract the image (C) that the territory, cloud sector is arranged in the cloud atlas picture should be arranged in the cloudless image relatively, with image (B) with have cloud atlas to look like to carry out computing, extract the image (D) that cloudless part in the cloud atlas picture is arranged;
E, image compound operation, with image (C) with (D) superpose computing, form a complete image (E) that goes behind the cloud;
Total some noise spot of the image that forms among f, the step e is removed noise spot by selecting smoothing filter, the image of step e is gone bad a little repaired again, gets last figure as a result (F).
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910113560A CN101710416A (en) | 2009-12-07 | 2009-12-07 | Processing method for multiple-target remote sensing image clouds |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910113560A CN101710416A (en) | 2009-12-07 | 2009-12-07 | Processing method for multiple-target remote sensing image clouds |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101710416A true CN101710416A (en) | 2010-05-19 |
Family
ID=42403201
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910113560A Pending CN101710416A (en) | 2009-12-07 | 2009-12-07 | Processing method for multiple-target remote sensing image clouds |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101710416A (en) |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103310430A (en) * | 2012-03-13 | 2013-09-18 | 三星电子株式会社 | Method and apparatus for deblurring non-uniform motion blur |
| CN103873758A (en) * | 2012-12-17 | 2014-06-18 | 北京三星通信技术研究有限公司 | Method, device and equipment for generating panorama in real time |
| CN104484859A (en) * | 2014-10-20 | 2015-04-01 | 电子科技大学 | Multispectral optical remote sensing image data thin-cloud removing method |
| CN104732503A (en) * | 2013-12-24 | 2015-06-24 | 中国科学院深圳先进技术研究院 | Image defogging and enhancement method and device |
| CN105574825A (en) * | 2015-12-16 | 2016-05-11 | 中国科学院遥感与数字地球研究所 | Method and device for removing clouds and vertical stripes of multispectral remote sensing images |
| CN105893977A (en) * | 2016-04-25 | 2016-08-24 | 福州大学 | Rice mapping method based on self-adaptive feature selection |
| CN106682419A (en) * | 2016-12-27 | 2017-05-17 | 深圳先进技术研究院 | Fitting method and device for medical image parameters |
| CN109029383A (en) * | 2018-05-03 | 2018-12-18 | 山东省科学院海洋仪器仪表研究所 | A kind of rationally distributed property evaluation method of oceanographic buoy erect-position based on spatial analysis |
| CN109166089A (en) * | 2018-07-24 | 2019-01-08 | 重庆三峡学院 | The method that a kind of pair of multispectral image and full-colour image are merged |
| CN109934788A (en) * | 2019-03-22 | 2019-06-25 | 鲁东大学 | A kind of remote sensing images missing data restorative procedure based on standard remote sensing images |
| CN115082452A (en) * | 2022-07-26 | 2022-09-20 | 北京数慧时空信息技术有限公司 | Cloud and shadow based quantitative evaluation method for quality of remote sensing image |
-
2009
- 2009-12-07 CN CN200910113560A patent/CN101710416A/en active Pending
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103310430A (en) * | 2012-03-13 | 2013-09-18 | 三星电子株式会社 | Method and apparatus for deblurring non-uniform motion blur |
| CN103873758A (en) * | 2012-12-17 | 2014-06-18 | 北京三星通信技术研究有限公司 | Method, device and equipment for generating panorama in real time |
| CN103873758B (en) * | 2012-12-17 | 2018-09-21 | 北京三星通信技术研究有限公司 | The method, apparatus and equipment that panorama sketch generates in real time |
| CN104732503A (en) * | 2013-12-24 | 2015-06-24 | 中国科学院深圳先进技术研究院 | Image defogging and enhancement method and device |
| CN104732503B (en) * | 2013-12-24 | 2017-10-24 | 中国科学院深圳先进技术研究院 | Image defogging Enhancement Method and device |
| CN104484859B (en) * | 2014-10-20 | 2017-09-01 | 电子科技大学 | A kind of method that multispectral remote sensing image data remove thin cloud |
| CN104484859A (en) * | 2014-10-20 | 2015-04-01 | 电子科技大学 | Multispectral optical remote sensing image data thin-cloud removing method |
| CN105574825A (en) * | 2015-12-16 | 2016-05-11 | 中国科学院遥感与数字地球研究所 | Method and device for removing clouds and vertical stripes of multispectral remote sensing images |
| CN105574825B (en) * | 2015-12-16 | 2018-08-10 | 中国科学院遥感与数字地球研究所 | A kind of method and apparatus that multi-spectral remote sensing image removes cloud and removes band |
| CN105893977A (en) * | 2016-04-25 | 2016-08-24 | 福州大学 | Rice mapping method based on self-adaptive feature selection |
| CN105893977B (en) * | 2016-04-25 | 2019-08-09 | 福州大学 | A kind of rice drafting method based on adaptive features select |
| CN106682419A (en) * | 2016-12-27 | 2017-05-17 | 深圳先进技术研究院 | Fitting method and device for medical image parameters |
| CN106682419B (en) * | 2016-12-27 | 2019-05-07 | 深圳先进技术研究院 | Fitting method and device for medical image parameters |
| CN109029383A (en) * | 2018-05-03 | 2018-12-18 | 山东省科学院海洋仪器仪表研究所 | A kind of rationally distributed property evaluation method of oceanographic buoy erect-position based on spatial analysis |
| CN109166089A (en) * | 2018-07-24 | 2019-01-08 | 重庆三峡学院 | The method that a kind of pair of multispectral image and full-colour image are merged |
| CN109934788A (en) * | 2019-03-22 | 2019-06-25 | 鲁东大学 | A kind of remote sensing images missing data restorative procedure based on standard remote sensing images |
| CN115082452A (en) * | 2022-07-26 | 2022-09-20 | 北京数慧时空信息技术有限公司 | Cloud and shadow based quantitative evaluation method for quality of remote sensing image |
| CN115082452B (en) * | 2022-07-26 | 2022-11-04 | 北京数慧时空信息技术有限公司 | Cloud and shadow based quantitative evaluation method for quality of remote sensing image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101710416A (en) | Processing method for multiple-target remote sensing image clouds | |
| WO2021120406A1 (en) | Infrared and visible light fusion method based on saliency map enhancement | |
| Huang et al. | Cloud removal from optical satellite imagery with SAR imagery using sparse representation | |
| CN111899194A (en) | Method for removing cloud and cloud shadow in remote sensing image | |
| CN109767454B (en) | Unmanned aerial vehicle aerial video moving target detection method based on time-space-frequency significance | |
| CN103049898A (en) | Method for fusing multispectral and full-color images with light cloud | |
| CN109523585B (en) | Multisource remote sensing image feature matching method based on direction phase consistency | |
| CN105405138B (en) | Waterborne target tracking based on conspicuousness detection | |
| JP6233869B2 (en) | Image processing apparatus, image processing apparatus control method, and program | |
| CN103218811A (en) | Statistical distribution-based satellite multi-spectral image waveband registration method | |
| CN102663385A (en) | Detection method for spot target on satellite | |
| CN107341809B (en) | Multi-source remote sensing image-oriented coastline automatic extraction and deviation correction method | |
| Jiang et al. | Unmixing-based spatiotemporal image fusion accounting for complex land cover changes | |
| CN112598709B (en) | Pedestrian movement speed intelligent sensing method based on video stream | |
| CN110688923A (en) | Sentinel 1A SAR data-based urban inland inundation risk area extraction method | |
| CN102629368B (en) | Color image vignetting recovery method based on illumination surface modeling | |
| CN106780414A (en) | A kind of processing method of multiple-target remote sensing image clouds | |
| Xiao et al. | A Novel Image Fusion Method for Water Body Extraction Based on Optimal Band Combination. | |
| CN108537807B (en) | Improved edge enhancement detection method based on high-resolution fourth-order satellite image | |
| Wendleder et al. | Water body detection from TanDEM-X data: Concept and first evaluation of an accurate water indication mask | |
| CN112734819B (en) | Earth surface filtering method suitable for high-resolution remote sensing satellite DSM | |
| CN105469421A (en) | Method based on panoramic system for achieving monitoring of ground moving target | |
| CN110136128B (en) | SAR image change detection method based on Rao detection | |
| Yang et al. | Optical Remote Sensing Image Optimized Dehazing Algorithm Based On Hot | |
| El-Mezouar et al. | A high-resolution index for vegetation extraction in IKONOS images |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Open date: 20100519 |