WO2015037189A1 - ノイズ除去装置、方法およびプログラム - Google Patents
ノイズ除去装置、方法およびプログラム Download PDFInfo
- Publication number
- WO2015037189A1 WO2015037189A1 PCT/JP2014/004265 JP2014004265W WO2015037189A1 WO 2015037189 A1 WO2015037189 A1 WO 2015037189A1 JP 2014004265 W JP2014004265 W JP 2014004265W WO 2015037189 A1 WO2015037189 A1 WO 2015037189A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- resolution
- decomposition
- component
- low
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 53
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 133
- 238000003384 imaging method Methods 0.000 claims description 8
- 230000001771 impaired effect Effects 0.000 abstract description 2
- 238000012935 Averaging Methods 0.000 description 5
- 230000000694 effects Effects 0.000 description 4
- 238000000149 argon plasma sintering Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000003672 processing method Methods 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N1/00—Scanning, transmission or reproduction of documents or the like, e.g. facsimile transmission; Details thereof
- H04N1/40—Picture signal circuits
- H04N1/409—Edge or detail enhancement; Noise or error suppression
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N25/00—Circuitry of solid-state image sensors [SSIS]; Control thereof
- H04N25/60—Noise processing, e.g. detecting, correcting, reducing or removing noise
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N25/00—Circuitry of solid-state image sensors [SSIS]; Control thereof
- H04N25/60—Noise processing, e.g. detecting, correcting, reducing or removing noise
- H04N25/618—Noise processing, e.g. detecting, correcting, reducing or removing noise for random or high-frequency noise
-
- 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
- G06T2207/10036—Multispectral image; Hyperspectral image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20016—Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
Definitions
- the present invention relates to a noise removal apparatus, method, and program.
- a system for photographing the ground from a satellite or an aircraft has been used.
- it is widely performed to simultaneously capture both a high-resolution panchromatic image and a multi-spectral image having a lower resolution than the panchromatic image for the same subject.
- imaging is performed for all visible and near infrared wavelengths
- the multispectral sensor imaging is performed only for specific wavelengths such as blue and red.
- an image of a specific color can be obtained, but the amount of light per unit area received by the sensor is smaller than that of a panchromatic sensor.
- the noise component increases with respect to the signal component, and a good quality image cannot be obtained.
- pan-sharpening processing is performed to create a high-resolution color image by combining both images.
- each pixel component of a multispectral image is decomposed into a luminance component, a hue component, and a saturation component, and the luminance component is replaced with a component of a high-resolution panchromatic image.
- a color image is created.
- removal of noise contained in an image is a general problem, not limited to high-resolution color images created by pan-sharpening processing, and such noise removal is performed by image processing.
- a method related to noise removal by image processing is described in Patent Document 1.
- Patent Document 1 The method described in Patent Document 1 is as follows.
- An image processing method for removing noise contained in an image has an image input procedure for inputting an original image composed of a plurality of pixels.
- the input original image is decomposed, a multi-resolution image generation procedure for sequentially generating a plurality of low-frequency images having a low resolution, a plurality of high-frequency images having a sequentially low resolution, a low-frequency image, A noise removal processing procedure for performing noise removal processing on each of the high-frequency images.
- an image acquisition procedure for obtaining an image from which noise has been removed from the original image based on the results of both the low-frequency image from which noise has been removed and the high-frequency image from which noise has been removed is provided.
- the original signal of an image may be damaged due to noise removal performed on one image.
- a general noise removal method including the method described in Patent Document 1, the nature of noise is preliminarily assumed, and noise components are estimated based on the noise property to remove noise.
- the component of the original signal of the image may be determined as noise and removed. After such noise removal has been applied to each of the panchromatic and multispectral images, the original signal components may be lost even in images created by pan-sharpening based on those images. Is concerned.
- noise removal is performed in noise removal of an image created by pan-sharpening processing or the like based on a plurality of images having different resolutions that are simultaneously imaged from the same photographing target. Due to the nature of the method, there is a problem that the original signal component of the image is impaired.
- An object of the present invention is to provide a noise removal apparatus, method, and program for solving this problem.
- the noise removal apparatus multiplexes an input first image having a predetermined resolution onto a component having a predetermined first resolution lower than the predetermined resolution of the input second image obtained by simultaneously imaging the same subject as the first image.
- image reconstruction means for reconstructing and outputting the first image.
- the noise removal method of the present invention multiplexes an input first image having a predetermined resolution onto a component having a predetermined first resolution that is lower than the predetermined resolution of the input second image obtained by simultaneously imaging the same subject as the first image. Resolving resolution, correcting at least one decomposition component created by the decomposition using the second image, replacing the decomposition component with a decomposition component created by the correction, and reconstructing the first image; It is characterized by outputting.
- the original signal component of an image can be lost in noise removal of an image created by pan-sharpening processing or the like based on a plurality of images with different resolutions simultaneously captured from the same subject. Can be reduced.
- FIG. 1 shows an example of the configuration of a noise removal apparatus according to this embodiment.
- the noise removal apparatus 10 includes a high resolution image input unit 11, a low resolution image input unit 12, a multiresolution decomposition unit 13, a low resolution component correction unit 14, a reconstruction unit 15, and an image output unit 16.
- the high resolution image input unit 11 outputs the input high resolution image.
- the low resolution image input unit 12 outputs the input low resolution image.
- the multi-resolution decomposition unit 13 performs multi-resolution decomposition that decomposes an input high-resolution image into low-resolution components.
- the low resolution component correction unit 14 corrects the low resolution component created by decomposing the high resolution image using the input low resolution image.
- the reconstruction unit 15 reconstructs a high-resolution image by replacing the corrected low-resolution component with the corrected low-resolution component.
- the image output unit 16 outputs a reconstructed high resolution image.
- the high resolution image input unit 11 receives a high resolution image to be imaged
- the low resolution image input unit 12 receives the same low resolution image as the high resolution image.
- the multi-resolution decomposition unit 13 multi-resolution decomposes the high-resolution image output from the high-resolution image input unit 11 into low-resolution components. At this time, the image is decomposed to a resolution corresponding to the input low-resolution image.
- the low resolution component correction unit 14 corrects a low resolution component (corresponding to an LL component, which will be described later) created by decomposing a high resolution image by the multiresolution decomposition unit 13 using the low resolution image output from the low resolution image input unit 12. To do.
- the reconstruction unit 15 replaces the low resolution component corrected by the low resolution component correction unit 14 in the multi-resolution decomposition unit 13 with the low resolution component corrected by the low resolution component correction unit 14 to reconstruct a high resolution image. To do.
- the image output unit 16 outputs the high resolution image reconstructed by the reconstruction unit 15.
- the high resolution image input unit 11 uses a high resolution CCD (Charge Coupled Device) camera
- the low resolution image input unit 12 uses a quarter resolution of the high resolution CCD camera (the length of one side of the target).
- Each image is input from a low-resolution CCD camera having one-fourth the number of pixels per pixel.
- the multi-resolution decomposition unit 13 performs multi-resolution decomposition by wavelet conversion.
- (1) is an input high resolution image
- (2) is a processed image.
- the input high resolution image is decomposed into four components.
- the upper left LL component is the average component in the vertical and horizontal directions
- the upper right HL component is the difference from the average in the vertical direction
- the lower left LH component is the difference from the average in the horizontal direction
- the lower right HH component is the difference from the average in the vertical and horizontal direction.
- the average component is a scaling coefficient
- the difference from the average is expressed by a wavelet coefficient. Note that the size of one pixel is the same in (1) and (2), and the length of one side of each component in (2) is half that of the input high-resolution image.
- (3) is an LL component in the previous processing
- (4) is an image that has been processed and decomposed into four components.
- the upper left LLLL component is the average component in the vertical and horizontal directions
- the upper right LLHL component is the difference from the average in the vertical direction
- the lower left LLLH component is the difference from the average in the horizontal direction
- the lower right HHHH component is the difference from the average in the vertical and horizontal direction. It is.
- the result is (5). Note that the size of one pixel is the same in (3), (4), and (5), and the length of one side of each component in (4) is 1/4 of the input high-resolution image. It has become.
- the correction in the low resolution component correction unit 14 will be described with reference to FIG.
- the low resolution component correction unit 14 corrects the low resolution component (LLLL component having the same resolution as that of the low resolution image) created by the high resolution image being decomposed by the multi-resolution decomposition unit 13.
- the low resolution component an average of the low resolution component and the low resolution image is used as the corrected low resolution component.
- Noise has a property of being randomly generated, and there is no correlation between noises included in the low resolution component and the low resolution image. Therefore, the noise component is reduced by averaging the corresponding pixels of the low resolution component and the low resolution image.
- the original signal of the image is included in both the low resolution component and the low resolution image, so that even if averaged, it is not reduced. Thus, only the noise component is reduced by this correction. In this correction, alignment processing is performed between the low resolution image and the high resolution image.
- the reconstruction unit 15 performs a reverse process of multi-resolution decomposition on the basis of the low resolution component created by decomposing and correcting the high resolution image, thereby reconstructing the high resolution image.
- the reconstructed high-resolution image is an image with reduced noise compared to the input high-resolution image.
- the alignment process is performed between the low-resolution image and the high-resolution image, so that the high-resolution image is properly reconstructed.
- the resolution of the low-resolution image is set to 1/4 of that of the high-resolution image
- the resolution ratio is not limited to this as long as it is less than 1/10.
- the multiresolution decomposition method is Wavelet transform
- the resolution is halved by one decomposition, so it is desirable that the resolution ratio be the reciprocal of a power of 2.
- the method of multi-resolution decomposition may be other than Wavelet transform, and the resolution ratio may be a reciprocal of a natural number if possible in other methods.
- the image may be interpolated.
- the high resolution CCD camera may be replaced with a panchromatic sensor
- the low resolution CCD camera may be replaced with a multispectral sensor.
- an average of the images of a plurality of bands of the multispectral sensor may be used as the low resolution image.
- images are taken in a plurality of wavelength bands. For example, when a multispectral sensor is composed of four bands of blue, green, red, and near infrared, and a panchromatic sensor captures an image in the visible wavelength range from blue to red, the multispectral sensor blue, green, A low resolution image may be obtained by averaging three red band images.
- the one having the best SN (Signal-to-Noise Ratio) characteristic may be set as the low resolution image, or the average of the plurality of images corresponding to the SN characteristic is set to the low resolution. It may be an image. If there is a positional shift between the images of the multispectral sensor, edges and contours may be blurred when a large number of images are averaged. Therefore, a low resolution image may be obtained by selecting and averaging several images with good SN characteristics to balance the effect of noise removal and the possibility of blurring of edges and contours.
- the average of the low resolution component and the low resolution image is used as the corrected low resolution component, but the average may be a weighted average or the like.
- the low resolution image may be used as a corrected low resolution component with less noise.
- the correction method is not limited to those described above, and any correction method may be used as long as the difference in noise between the low resolution image and the high resolution image is used.
- the original signal component may be lost in noise removal of images created by pan-sharpening processing, etc., based on a plurality of images with different resolutions that are captured simultaneously from the same subject. Can be reduced.
- the present embodiment is different from the first embodiment in that it includes a low-resolution image multi-resolution decomposition unit 27 and the operations of the multi-resolution decomposition unit and the low-resolution component correction unit.
- a low-resolution image multi-resolution decomposition unit 27 and the operations of the multi-resolution decomposition unit and the low-resolution component correction unit.
- different parts will be mainly described.
- the low resolution image multi-resolution decomposition unit 27 multi-resolution decomposes the input low resolution image into low resolution components.
- the multi-resolution decomposition unit 23 multi-resolution decomposes the input high-resolution image up to the resolution of the low-resolution component created by decomposing the low-resolution image.
- the low resolution component correction unit 24 corrects the low resolution component created by decomposing the high resolution image using the low resolution component created by decomposing the low resolution image.
- the correction in the low resolution component correction unit 24 will be described with reference to FIG. Assume that the resolution of the low-resolution image is half that of the high-resolution image, and the multi-resolution decomposition is performed for each of the low-resolution image and the high-resolution image up to half the resolution of the low-resolution image.
- the low resolution image is decomposed into LL, HL, LH, and HH components, and the LL component created by decomposing the high resolution image is decomposed into LLLL, LLHL, LLLH, and LLHH components.
- the low resolution component correction unit 24 corrects the LLLL, LLHL, LLLH, and LLHH components of the high resolution image with the LL, HL, LH, and HH components of the low resolution image, respectively. Regarding the correction, an average of the decomposition component of the high resolution image and the decomposition component of the low resolution image is used as the corrected decomposition component.
- the resolution of the low-resolution image is half that of the high-resolution image
- the resolution ratio is not limited to this as long as it is less than one-tenth.
- the multiresolution decomposition method is Wavelet transform
- the resolution is halved by one decomposition, so it is desirable that the resolution ratio be the reciprocal of a power of 2.
- the method of multi-resolution decomposition may be other than Wavelet transform, and the resolution ratio may be a reciprocal of a natural number if possible in other methods.
- the image may be interpolated.
- a high-resolution CCD camera and a low-resolution CCD camera, or a panchromatic sensor and a multispectral sensor may be used for image input.
- a multispectral sensor an average of the images of a plurality of bands of the multispectral sensor may be used as a low resolution image.
- an image having the best SN characteristic may be a low resolution image, and an average of a plurality of images corresponding to the SN characteristic may be a low resolution image.
- the average of the decomposition components of the high-resolution image and the low-resolution image is used as the corrected decomposition component, but the average may be a weighted average or the high-resolution image or low-resolution image.
- the image decomposition component may be a corrected decomposition component.
- the method of correcting the decomposition component may be different for each decomposition component having different L / H. For example, when it is known that noise at a certain frequency is included only in the high-resolution image because it is easy to get noise in the high-resolution image, the decomposition component of the low-resolution image is corrected for the decomposition component corresponding to that frequency. It may be a decomposed component. Further, according to the frequency characteristics of the high resolution image and the low resolution image, the decomposition components of the high resolution image and the low resolution image may be corrected by weighted averaging with different weights for each decomposition component. As a result, the effect of noise removal can be enhanced.
- the correction method is not limited to those described above, and any correction method may be used as long as the difference in noise between the low resolution image and the high resolution image is used.
- Such multi-resolution decomposition and correction may be further performed on the decomposition component corresponding to the LL component.
- the original signal component may be lost in noise removal of images created by pan-sharpening processing, etc., based on a plurality of images with different resolutions that are captured simultaneously from the same subject. Can be reduced.
- the present embodiment includes a plurality of low-resolution image input units 32 and a low-resolution image multi-resolution decomposition unit 37, and the operation of the low-resolution component correction unit is the same as that of the second embodiment. Different. Here, different parts will be mainly described.
- the low resolution image input unit 32 outputs the input low resolution image.
- the low-resolution image multi-resolution decomposition unit 37 performs multi-resolution decomposition that divides each input low-resolution image into low-resolution components.
- the low resolution component correction unit 34 corrects a low resolution component created by decomposing a high resolution image using each low resolution component created by decomposing a plurality of low resolution images.
- the high resolution image input unit 31 includes a high resolution panchromatic sensor
- the low resolution image input units 32-1 to 32-4 include, for example, a blue band, a green band, a red band, and a near infrared band multispectral sensor.
- An image is input.
- the low-resolution image multi-resolution decomposition units 37-1 to 37-4 perform multi-resolution decomposition up to a predetermined resolution (for example, half of the low-resolution image) for the input low-resolution images 1 to 4 of each band.
- the multi-resolution decomposition unit 33 performs multi-resolution decomposition up to the same predetermined resolution.
- the low resolution component correction unit 33 sets the average of the resolution component of the high resolution image and the resolution components of the low resolution images 1 to 4 as the corrected resolution component.
- the averaging is, for example, a weighted average, and weighting is performed with resolution and band characteristics.
- a band with a short wavelength, such as a blue band has a tendency to increase noise at a high frequency compared to a band with a long wavelength, such as a red band or a near-infrared band, because light scattering in the atmosphere increases. . For this reason, weighting is performed so as to increase the weighting of a band having a longer wavelength as the resolution is higher.
- the number of the low resolution image input unit and the low resolution image multi-resolution decomposition unit is four, it is not limited to this number.
- the original signal component may be lost in noise removal of images created by pan-sharpening processing, etc., based on a plurality of images with different resolutions that are captured simultaneously from the same subject. Can be reduced.
- FIG. 9 shows an example of the configuration of the noise removal apparatus in the present embodiment.
- the noise removing device 40 includes a multi-resolution decomposition unit 41, a low-resolution component correction unit 42, and a reconstruction unit 43.
- a first image having a predetermined first resolution is input to the multi-resolution decomposition unit 41.
- a second image having a predetermined second resolution lower than the first resolution obtained by simultaneously capturing the same subject as the first image is input to the low resolution component correction unit 42.
- the multi-resolution decomposition unit 41 decomposes the first image into components of the second resolution.
- the low resolution component correction unit 42 corrects at least one low resolution component created by the decomposition using the second image.
- the reconstruction unit 43 replaces the low resolution component with the low resolution component created by the correction, reconstructs the first image, and outputs it.
- the original signal component may be lost in noise removal of images created by pan-sharpening processing, etc., based on a plurality of images with different resolutions that are captured simultaneously from the same subject. Can be reduced.
- the noise removal device according to Supplementary Note 1, wherein the noise removal device is characterized in that (Supplementary note 3) The supplementary method 1 or 2, wherein the correction method includes a weighted average of the decomposition component of the first image and the decomposition component of the second image or the second image.
- Noise removal device (Supplementary note 4) The noise removing device according to supplementary notes 1 to 3, wherein the second image is a plurality of images having different characteristics. (Supplementary note 5) The noise removing device according to supplementary notes 1 to 4, wherein the first image is a panchromatic image, and the second image is a multispectral image.
- the input first image having a predetermined resolution is subjected to multi-resolution decomposition into a component having a predetermined first resolution lower than the predetermined resolution of the second image input by simultaneously imaging the same subject as the first image, Correcting at least one decomposition component created by the decomposition using the second image; A noise removal method, wherein the decomposition component is replaced with a decomposition component created by the correction, and the first image is reconstructed and output.
- the second image is subjected to multi-resolution decomposition into a predetermined second resolution component lower than the first resolution
- the first image is subjected to multiresolution decomposition into the second resolution component
- the at least one decomposition component created by decomposing the first image into the second resolution is corrected using a decomposition component created by decomposing the corresponding second image.
- the noise removal method according to appendix 6, characterized by: (Supplementary note 8) The supplementary method 6 or 7, wherein the correction method includes a weighted average of the decomposition component of the first image and the decomposition component of the second image or the second image. Noise removal method.
- a low-resolution image multi-resolution decomposition step that multi-resolution decomposes the second image into components of a predetermined second resolution lower than the first resolution,
- the multi-resolution decomposition step multi-resolution decomposes the first image into components of the second resolution;
- the low-resolution component correction step at least one decomposition component created by decomposing the first image into the second resolution is corrected using a decomposition component created by decomposing the corresponding second image.
- a storage medium for storing the removal program. (Supplementary note 13) The storage medium for storing the noise removal program according to supplementary notes 10 to 12, wherein the second image is a plurality of images having different characteristics.
- the present invention can be used in a noise removal apparatus, method, and program.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
- Studio Devices (AREA)
Abstract
Description
(付記1)入力された所定解像度の第1画像を前記第1画像と同一の被写体が同時に撮像され入力された第2画像の前記所定解像度より低い所定第1解像度の成分に多重解像度分解する多重解像度分解手段と、
前記分解により作成された少なくとも1つの分解成分について前記第2画像を用いて補正する低解像度成分補正手段と、
前記分解成分を前記補正により作成された分解成分で置き換え、前記第1画像を再構成し、出力する画像再構成手段と
を備えることを特徴とするノイズ除去装置。
(付記2)前記第2画像を前記第1解像度より低い所定第2解像度の成分に多重解像度分解する低解像度画像多重解像度分解手段をさらに備え、
前記多重解像度分解手段は、前記第1画像を前記第2解像度の成分に多重解像度分解し、
前記低解像度成分補正手段は、前記第1画像の第2解像度への分解により作成された少なくとも1つの分解成分を対応する前記第2画像の分解により作成された分解成分を用いて補正する
ことを特徴とする付記1に記載のノイズ除去装置。
(付記3)前記補正の方法は、前記第1画像の分解成分と前記第2画像または前記第2画像の分解成分とを加重平均することを含むことを特徴とする付記1または2に記載のノイズ除去装置。
(付記4)前記第2画像は、特性の異なった複数の画像であることを特徴とする付記1乃至3に記載のノイズ除去装置。
(付記5)前記第1画像はパンクロマチック画像、前記第2画像はマルチスペクトル画像であることを特徴とする付記1乃至4に記載のノイズ除去装置。
(付記6)入力された所定解像度の第1画像を前記第1画像と同一の被写体が同時に撮像され入力された第2画像の前記所定解像度より低い所定第1解像度の成分に多重解像度分解し、
前記分解により作成された少なくとも1つの分解成分について前記第2画像を用いて補正し、
前記分解成分を前記補正により作成された分解成分で置き換え、前記第1画像を再構成し、出力する
ことを特徴とするノイズ除去方法。
(付記7)さらに、前記第2画像を前記第1解像度より低い所定第2解像度の成分に多重解像度分解し、
前記多重解像度分解する際は、前記第1画像を前記第2解像度の成分に多重解像度分解し、
前記低解像度成分補正する際は、前記第1画像の第2解像度への分解により作成された少なくとも1つの分解成分を対応する前記第2画像の分解により作成された分解成分を用いて補正する
ことを特徴とする付記6に記載のノイズ除去方法。
(付記8)前記補正の方法は、前記第1画像の分解成分と前記第2画像または前記第2画像の分解成分とを加重平均することを含むことを特徴とする付記6または7に記載のノイズ除去方法。
(付記9)前記第2画像は特性の異なった複数の画像であることを特徴とする付記6乃至8に記載のノイズ除去方法。
(付記10)ノイズ除去装置を構成するコンピュータに、
入力された所定解像度の第1画像を前記第1画像と同一の被写体が同時に撮像され入力された第2画像の前記所定解像度より低い所定第1解像度の成分に多重解像度分解する多重解像度分解ステップと、
前記分解により作成された少なくとも1つの分解成分について前記第2画像を用いて補正する低解像度成分補正ステップと、
前記分解成分を前記補正により作成された分解成分で置き換え、前記第1画像を再構成し、出力する画像再構成ステップと
を実行させることを特徴とするノイズ除去プログラムを格納する記憶媒体。
(付記11)前記第2画像を前記第1解像度より低い所定第2解像度の成分に多重解像度分解する低解像度画像多重解像度分解ステップをさらに備え、
前記多重解像度分解ステップは、前記第1画像を前記第2解像度の成分に多重解像度分解し、
前記低解像度成分補正ステップは、前記第1画像の第2解像度への分解により作成された少なくとも1つの分解成分を対応する前記第2画像の分解により作成された分解成分を用いて補正する
ことを特徴とする付記10に記載のノイズ除去プログラム。
(付記12)前記補正方法は、前記第1画像の分解成分と前記第2画像または前記第2画像の分解成分とを加重平均することを含むことを特徴とする付記10または11に記載のノイズ除去プログラムを格納する記憶媒体。
(付記13)前記第2画像は特性の異なった複数の画像であることを特徴とする付記10乃至12に記載のノイズ除去プログラムを格納する記憶媒体。
11、21、31 高解像度画像入力部
12、22、32-1~4 低解像度画像入力部
13、23、33、41 多重解像度分解部
14、24、34、42 低解像度成分補正部
15、25、35、43 再構成部
16、26、36 画像出力部
27、37-1~4 低解像度画像多重解像度分解部
Claims (10)
- 入力された所定解像度の第1画像を前記第1画像と同一の被写体が同時に撮像され入力された第2画像の前記所定解像度より低い所定第1解像度の成分に多重解像度分解する多重解像度分解手段と、
前記分解により作成された少なくとも1つの分解成分について前記第2画像を用いて補正する低解像度成分補正手段と、
前記分解成分を前記補正により作成された分解成分で置き換え、前記第1画像を再構成し、出力する画像再構成手段と
を備えることを特徴とするノイズ除去装置。 - 前記第2画像を前記第1解像度より低い所定第2解像度の成分に多重解像度分解する低解像度画像多重解像度分解手段をさらに備え、
前記多重解像度分解手段は、前記第1画像を前記第2解像度の成分に多重解像度分解し、
前記低解像度成分補正手段は、前記第1画像の第2解像度への分解により作成された少なくとも1つの分解成分を対応する前記第2画像の分解により作成された分解成分を用いて補正する
ことを特徴とする請求項1に記載のノイズ除去装置。 - 前記補正の方法は、前記第1画像の分解成分と前記第2画像または前記第2画像の分解成分とを加重平均することを含むことを特徴とする請求項1または2に記載のノイズ除去装置。
- 前記第2画像は、特性の異なった複数の画像であることを特徴とする請求項1乃至3に記載のノイズ除去装置。
- 前記第1画像はパンクロマチック画像、前記第2画像はマルチスペクトル画像である
ことを特徴とする請求項1乃至4に記載のノイズ除去装置。 - 入力された所定解像度の第1画像を前記第1画像と同一の被写体が同時に撮像され入力された第2画像の前記所定解像度より低い所定第1解像度の成分に多重解像度分解し、
前記分解により作成された少なくとも1つの分解成分について前記第2画像を用いて補正し、
前記分解成分を前記補正により作成された分解成分で置き換え、前記第1画像を再構成し、出力する
ことを特徴とするノイズ除去方法。 - さらに、前記第2画像を前記第1解像度より低い所定第2解像度の成分に多重解像度分解し、
前記多重解像度分解する際は、前記第1画像を前記第2解像度の成分に多重解像度分解し、
前記低解像度成分補正する際は、前記第1画像の第2解像度への分解により作成された少なくとも1つの分解成分を対応する前記第2画像の分解により作成された分解成分を用いて補正する
ことを特徴とする請求項6に記載のノイズ除去方法。 - 前記補正の方法は、前記第1画像の分解成分と前記第2画像または前記第2画像の分解成分とを加重平均することを含むことを特徴とする請求項6または7に記載のノイズ除去方法。
- 前記第2画像は特性の異なった複数の画像であることを特徴とする請求項6乃至8に記載のノイズ除去方法。
- ノイズ除去装置を構成するコンピュータに、
入力された所定解像度の第1画像を前記第1画像と同一の被写体が同時に撮像され入力された第2画像の前記所定解像度より低い所定第1解像度の成分に多重解像度分解する多重解像度分解ステップと、
前記分解により作成された少なくとも1つの分解成分について前記第2画像を用いて補正する低解像度成分補正ステップと、
前記分解成分を前記補正により作成された分解成分で置き換え、前記第1画像を再構成し、出力する画像再構成ステップと
を実行させることを特徴とするノイズ除去プログラムを格納する記憶媒体。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201480050517.XA CN105593898B (zh) | 2013-09-12 | 2014-08-20 | 降噪装置、方法及其程序 |
JP2015536437A JP6350532B2 (ja) | 2013-09-12 | 2014-08-20 | ノイズ除去装置、方法およびプログラム |
US14/912,992 US9972072B2 (en) | 2013-09-12 | 2014-08-20 | Noise reduction apparatus, method and program for the same |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2013-189549 | 2013-09-12 | ||
JP2013189549 | 2013-09-12 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2015037189A1 true WO2015037189A1 (ja) | 2015-03-19 |
Family
ID=52665322
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2014/004265 WO2015037189A1 (ja) | 2013-09-12 | 2014-08-20 | ノイズ除去装置、方法およびプログラム |
Country Status (4)
Country | Link |
---|---|
US (1) | US9972072B2 (ja) |
JP (1) | JP6350532B2 (ja) |
CN (1) | CN105593898B (ja) |
WO (1) | WO2015037189A1 (ja) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE112019007327T5 (de) | 2019-05-16 | 2022-03-24 | Mitsubishi Electric Corporation | Bildverarbeitungseinrichtung, Verfahren, Bildleseeinrichtung, Programm und Aufzeichnungsmedium |
US12026854B2 (en) | 2019-05-16 | 2024-07-02 | Mitsubishi Electric Corporation | Image processing device, method, image reading device, and recording medium |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07319092A (ja) * | 1994-03-31 | 1995-12-08 | Fuji Photo Film Co Ltd | 画像重ね合せ方法およびエネルギーサブトラクション方法 |
US20040141659A1 (en) * | 2003-01-17 | 2004-07-22 | Yun Zhang | System and method for image fusion |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0677780B1 (en) * | 1994-03-31 | 2003-05-28 | Fuji Photo Film Co., Ltd. | Image superposition processing method |
US5949914A (en) * | 1997-03-17 | 1999-09-07 | Space Imaging Lp | Enhancing the resolution of multi-spectral image data with panchromatic image data using super resolution pan-sharpening |
JP2005210441A (ja) * | 2004-01-23 | 2005-08-04 | Sanyo Electric Co Ltd | 画像信号処理装置 |
WO2007116543A1 (ja) | 2006-03-31 | 2007-10-18 | Nikon Corporation | 画像処理方法 |
JP4987688B2 (ja) * | 2007-12-25 | 2012-07-25 | 株式会社東芝 | 画像高解像度化方法および装置 |
JP5206466B2 (ja) * | 2009-02-13 | 2013-06-12 | 富士通株式会社 | 画像補正装置、画像補正プログラムおよび画像撮影装置 |
-
2014
- 2014-08-20 CN CN201480050517.XA patent/CN105593898B/zh active Active
- 2014-08-20 US US14/912,992 patent/US9972072B2/en active Active
- 2014-08-20 JP JP2015536437A patent/JP6350532B2/ja active Active
- 2014-08-20 WO PCT/JP2014/004265 patent/WO2015037189A1/ja active Application Filing
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH07319092A (ja) * | 1994-03-31 | 1995-12-08 | Fuji Photo Film Co Ltd | 画像重ね合せ方法およびエネルギーサブトラクション方法 |
US20040141659A1 (en) * | 2003-01-17 | 2004-07-22 | Yun Zhang | System and method for image fusion |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE112019007327T5 (de) | 2019-05-16 | 2022-03-24 | Mitsubishi Electric Corporation | Bildverarbeitungseinrichtung, Verfahren, Bildleseeinrichtung, Programm und Aufzeichnungsmedium |
US12026854B2 (en) | 2019-05-16 | 2024-07-02 | Mitsubishi Electric Corporation | Image processing device, method, image reading device, and recording medium |
Also Published As
Publication number | Publication date |
---|---|
US20160210729A1 (en) | 2016-07-21 |
CN105593898A (zh) | 2016-05-18 |
CN105593898B (zh) | 2018-10-12 |
JP6350532B2 (ja) | 2018-07-04 |
JPWO2015037189A1 (ja) | 2017-03-02 |
US9972072B2 (en) | 2018-05-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
KR101643319B1 (ko) | 하이 다이나믹 레인지 영상을 획득하는 장치 및 그 방법 | |
US10778917B2 (en) | Joint dictionary generation method for image processing, interlace-based high dynamic range imaging apparatus using joint dictionaries and image processing method of the same | |
JP5668105B2 (ja) | 画像処理装置、画像処理方法及び画像処理プログラム | |
JP6104985B2 (ja) | デジタル画像センサ、画像獲得方法及び画像再構成方法並びにその実行システム | |
US11423514B2 (en) | Image processing noise reduction | |
EP3058549A1 (en) | Converting an image from a dual-band sensor to a visible color image | |
KR20110091117A (ko) | 영상잡음 제거장치 및 그 방법 | |
JP2012533961A (ja) | 画質を向上させる方法 | |
JP2010531560A5 (ja) | ||
JP6350532B2 (ja) | ノイズ除去装置、方法およびプログラム | |
WO2019237659A1 (en) | Blind compressive sampling method, apparatus and imaging system | |
JP6595161B2 (ja) | 画像処理装置、画像処理方法、及び、撮像装置 | |
Matsuo et al. | Super-resolution for 2K/8K television using wavelet-based image registration | |
US20110261236A1 (en) | Image processing apparatus, method, and recording medium | |
JP5682443B2 (ja) | 画像処理装置、画像処理方法及び画像処理プログラム | |
JP4866756B2 (ja) | 撮像装置 | |
Luo et al. | Color image restoration via extended joint sparse model | |
Vega et al. | A Bayesian super-resolution approach to demosaicing of blurred images | |
Matsuo et al. | Super-resolution method by registration of multi-scale components considering color sampling pattern and frequency spectrum power of UHDTV camera | |
Guicquero et al. | On adaptive pixel random selection for compressive sensing | |
Sato et al. | Compressive color sensing using random complementary color filter array | |
Jayachandran et al. | Efficient demosaicking of color images using theoretic reconstruction technique | |
Buntilov et al. | Assessment of the method-inherent distortions in wavelet fusion | |
Huijuan et al. | Modulation transfer function online compensation of imaging system in remote sensing satellite | |
CN115719311A (zh) | 处理图像的方法和相关设备 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 14843619 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 14912992 Country of ref document: US |
|
ENP | Entry into the national phase |
Ref document number: 2015536437 Country of ref document: JP Kind code of ref document: A |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 14843619 Country of ref document: EP Kind code of ref document: A1 |