CN112017127A - Image strip noise removing method and system based on spectral filtering - Google Patents
Image strip noise removing method and system based on spectral filtering Download PDFInfo
- Publication number
- CN112017127A CN112017127A CN202010850347.0A CN202010850347A CN112017127A CN 112017127 A CN112017127 A CN 112017127A CN 202010850347 A CN202010850347 A CN 202010850347A CN 112017127 A CN112017127 A CN 112017127A
- Authority
- CN
- China
- Prior art keywords
- noise
- unit
- image
- detection unit
- column
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 66
- 238000001914 filtration Methods 0.000 title claims abstract description 38
- 230000003595 spectral effect Effects 0.000 title claims abstract description 15
- 238000001514 detection method Methods 0.000 claims abstract description 92
- 230000004044 response Effects 0.000 claims abstract description 78
- 230000000737 periodic effect Effects 0.000 claims abstract description 59
- 238000012935 Averaging Methods 0.000 claims abstract description 11
- 238000004364 calculation method Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 abstract description 15
- 230000008569 process Effects 0.000 abstract description 6
- 230000001788 irregular Effects 0.000 abstract description 3
- 239000000523 sample Substances 0.000 description 15
- 238000009499 grossing Methods 0.000 description 7
- 238000011156 evaluation Methods 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000011160 research Methods 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 230000005540 biological transmission Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000032683 aging Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 238000005316 response function Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Images
Classifications
-
- G06T5/70—
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/10—Image enhancement or restoration by non-spatial domain filtering
-
- 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
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
- G06T2207/10041—Panchromatic 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30204—Marker
Abstract
The invention relates to a method and a system for removing image strip noise based on spectral filtering, which are used for reading a panchromatic or multispectral image; identifying periodic noise and bad line noise in the image; compensating the periodic noise by adopting a gain adjustment method; and compensating the bad line noise by adopting a left-right column averaging method. The invention can realize the determination of the type of the strip noise and the rapid determination of the position of the strip noise by defining and calculating indexes such as the response capability of the detection unit, the response difference of the detection unit and the like of the image, the whole processing process is automatically realized by using a computer, the manual intervention is not needed, the processing speed is high, the automation degree is high, and the automatic batch processing of mass images can be realized. The method is suitable for satellite images with irregular stripe noise, and can be used for quickly identifying and removing the stripe noise positions and types of different types of remote sensing images such as panchromatic images, multispectrals images, hyperspectral images and the like.
Description
Technical Field
The invention relates to the technical field of remote sensing image processing, in particular to a method and a system for removing image strip noise based on spectral filtering.
Background
The remote sensing images are often subject to interference and influence from various noise sources during generation and transmission. Among them, the stripe noise is the most dominant one. The stripe noise is a special noise having a certain periodicity and directivity and being distributed in a stripe shape in the image.
The generation of stripe noise is mainly influenced by the imaging mechanism of the sensor, the factors of the instrument and the external factors. The essential reason is that longitudinal stripes with alternate light and shade appear on the image due to the non-identical response capability of each detection unit of the detector to the received signal, which is also the root cause of the generation of stripe noise. Specifically, the influencing factors of the stripe noise include:
(1) inhomogeneity of the detector device itself. Due to the limitation of the current material manufacturing and process, various defects such as uneven doping, unequal thickness, uneven size of detection elements and the like can occur in the material of the CCD device, so that parameters of different detection units are different, and the difference of response of the detection units is caused. In addition, wear and tear on the instrument, aging, etc. can also cause non-uniformity in the response of the detector.
(2) Influence of CCD dark current. The magnitude of the dark current is related to the intrinsic material, the magnitude, the temperature, etc. of the CCD pixels.
(3) Non-uniformity introduced by the working state. The driving signal of the CCD focal plane detector, the light energy of the imaging system and the like are related to the working state of the sensor, and the change of the working state influences the response consistency of the CCD focal plane detection unit of the whole sensor.
(4) Influence of external factors. In the working process, due to the interference of various external environments such as space electromagnetism and the like, the movement of the sensor, the difference of front and back response of the detection unit and the like, the working parameters and the working efficiency of the pixel of the focal plane device array can be influenced, and the inconsistency of the output signals of the device is caused.
(5) Incompleteness and incorrectness of CCD response function correction.
The existence of stripe noise causes problems in qualitative and quantitative use of the remote sensing image, and the stripe noise seriously influences the identification and extraction of the type of the ground object by an operator in qualitative remote sensing interpretation and interpretation. When a quantitative remote sensing product is inverted, the radiation information is distorted due to the stripe noise, the inversion result cannot be evaluated, and the use effect of the remote sensing image is influenced.
At present, research on stripe noise at home and abroad mainly focuses on a stripe noise removal method. The methods of stripe noise removal can be basically classified into two categories: namely, the spatial domain and transform domain banding noise removal.
The spatial domain method is to utilize a sliding window with a certain size and shape and different weighting coefficients to perform sliding filtering of an image spatial domain in an original image so as to remove stripe noise in the image, and common methods include a histogram matching method, a moment matching method and the like;
the transform domain method firstly adopts a certain image transform means to transform the image into a corresponding frequency domain, carries out certain denoising processing by analyzing the distribution condition of signals and noise in the frequency domain, and then transforms the signal and the noise into a space domain through inverse transform to realize the removal of the strip noise, and the common methods comprise a wavelet transform method, a Fourier transform method and the like.
However, in the current remote sensing image, the stripe noise sometimes does not exist all the time, and the generalized stripe noise includes the bad line noise of a single line and the stripe noise which is continuous periodically. The existing studies are still insufficient in the following two aspects.
(1) The existing research aims at stripe noise with certain regularity. However, at present, stripe noise of part of satellites appears randomly, and does not exist on each satellite image, so that no obvious rule exists.
(2) The existing research mainly focuses on the research of a noise removal method, the position judgment of noise and the analysis of noise types are mainly carried out in a manual mode, and the full-automatic identification is difficult to realize.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides the method and the system for removing the image strip noise based on the spectral filtering, the whole processing process is automatically realized by a computer without manual intervention, the processing speed is high, the automation degree is high, and the automatic batch processing of mass images can be realized.
In order to achieve the above object, the present invention provides a method for removing image stripe noise based on spectral filtering, which includes:
reading a panchromatic or multispectral image;
calculating the response capability of each detection unit in the image;
filtering the response capability of each detection unit;
calculating the variation difference value of the response capability of each detection unit after filtering, and marking the column if the variation difference value is larger than a set threshold value;
if no marked column exists, no compensation is carried out, and the image is output; if a marked detection unit is present:
calculating the difference between the marked row serial numbers, if the difference between the row serial numbers is smaller than a threshold, determining that the row serial numbers are periodic noise, and compensating the periodic noise of the row by adopting a gain adjustment method; otherwise, judging the bad line noise, and compensating the bad line noise of the row by adopting a left-right row averaging method.
Further, calculating the response capability of each detection unit in the image comprises: calculating the mean value D of the gray values of each column of the imageiAnd the mean value mu of the gray value of the whole image; calculating the response capability N of the ith detection uniti＝Di/μ。
Further, the filtering includes: setting a filter window by taking the ith detection unit as a center, wherein the size of the filter window is 2k +1, and calculating the average value of the response capability of each detection unit in the filter window as the filtered response capability N 'of the ith detection unit'i. Further, the variation difference of the ith detection unit is Ci＝N′i-NiThe set threshold is 5.
Further, compensating the column of periodic noise using a gain adjustment method, comprising:
calculating the normalized gain G of the ith detection uniti＝N′i/Ni；
Calculating the gray value mean value D after the i-th column compensationcal_i＝DiGi；
Adopting a left-right column average method to compensate the bad line noise of the column, comprising the following steps:
calculating the gray value mean value D 'of the compensated i-th column'i＝(Di-1+Di+1)/2。
Further, only the start column and the end column are marked for periodic noise; the column where the noise is located is marked for bad line noise.
The invention provides an image strip noise removing system based on spectral filtering, which comprises an image reading unit, a response capability calculating unit, a filtering unit, a difference marking unit, a judging unit, a periodic noise compensating unit, a bad line noise compensating unit and an output unit, wherein the image reading unit is used for reading the image strip noise;
the image reading unit reads a panchromatic or multispectral image;
the response capability calculation unit calculates the response capability of each detection unit in the image;
the filtering unit filters the response capability of each detection unit;
the difference marking unit marks the row if the variation difference of the response capability of each detection unit after filtering is calculated and is greater than a set threshold;
the judging unit judges, if no marked column exists, compensation is not carried out, and the output unit directly outputs the image; if the marked detection unit exists, calculating the difference between the marked row serial numbers, if the difference between the row serial numbers is smaller than a threshold, judging that the periodic noise exists, and outputting the starting row and the ending row of each group of the marked periodic noise to a periodic noise compensation unit; if the difference value of the serial numbers of the columns is not less than the threshold, the noise of the bad line is judged, and the marked columns are output to a noise compensation unit of the bad line;
the periodic noise compensation unit compensates each column covered by the initial column and the final column by adopting a gain adjustment method;
the bad line noise compensation unit compensates the marked column by adopting a left-right column averaging method;
the output unit outputs the image compensated by the periodic noise compensation unit and the bad line noise compensation unit.
Further, the response capability calculation unit calculates the response of each detection unit in the imageThe capabilities include: calculating the mean value D of the gray values of each column of the imageiAnd the mean value mu of the gray value of the whole image; calculating the response capability N of the ith detection uniti＝Di/μ。
Further, the filtering unit filters the response capability of each detection unit, and includes: receiving the response capability N of the ith detection unit output by the response capability calculation unitiSetting a filter window with the ith detection unit as the center, wherein the size of the filter window is 2k +1, and calculating the average value of the response capability of each detection unit in the filter window as the filtered response capability N 'of the ith detection unit'i。
Further, the difference marking unit calculates the variation difference of the ith detection unit as Ci＝N′i-NiAnd C isiCompared with a set threshold, the set threshold is 5.
Further, the periodic noise compensation unit compensates each column covered by the start column and the end column by using a gain adjustment method, and includes: calculating the normalized gain G of the ith detection uniti＝N′i/Ni(ii) a Calculating the gray value mean value D after the i-th column compensationcal_i＝DiGi。
Further, the bad line noise compensation unit compensates the marked column by adopting a left-right column averaging method, and comprises the following steps: calculating the gray value mean value D 'of the compensated i-th column'i＝(Di-1+Di+1)/2。
The technical scheme of the invention has the following beneficial technical effects:
(1) the invention can realize the determination of the type of the strip noise and the rapid determination of the position of the strip noise by defining and calculating indexes such as the response capability of the detection unit, the response difference of the detection unit and the like of the image, the whole processing process is automatically realized by using a computer, the manual intervention is not needed, the processing speed is high, the automation degree is high, and the automatic batch processing of mass images can be realized.
(2) The method can realize the rapid automatic identification of the periodic stripe noise and the bad line noise, and can not automatically identify when the image has no stripe noise; when the image has the stripe noise, the position of the stripe noise is automatically judged, and the removing algorithm of the stripe noise is automatically selected to realize the automatic removing of the stripe noise.
(3) The method is suitable for satellite images with irregular stripe noise, and can be used for quickly identifying and removing the stripe noise positions and types of different types of remote sensing images such as panchromatic images, multispectrals images, hyperspectral images and the like.
(4) The invention automatically identifies the periodic noise and the bad line noise, discriminates and compensates according to the characteristics of each noise, automatically selects a matched noise compensation algorithm according to the characteristics of the noise while keeping the original image information as much as possible, reduces or eliminates errors caused by inconsistent sensor probe pixels and the noise, can obviously improve the image quality and has good compensation effect.
Drawings
FIG. 1 is a schematic flow chart of noise removal;
FIG. 2 is a graph of bad line and periodic stripe noise;
FIG. 3 is a graph of response evaluation indexes of a detection unit before and after smoothing filtering;
FIG. 4 is an image after bad line and periodic stripe noise removal;
fig. 5 is a schematic diagram of a noise removal system.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the accompanying drawings in conjunction with the following detailed description. It should be understood that the description is intended to be exemplary only, and is not intended to limit the scope of the present invention. Moreover, in the following description, descriptions of well-known structures and techniques are omitted so as to not unnecessarily obscure the concepts of the present invention.
Aiming at the problem that a remote sensing image is often interfered and influenced by various noise sources in the generation and transmission processes, the invention provides an image strip noise removing method based on spectral filtering, and noise compensation is carried out on different noises. With reference to fig. 1, the method comprises the following steps:
step 1: reading of panchromatic or multispectral images.
Reading the panchromatic or multispectral image. It should be noted that the read image may have bad line noise or periodic stripe noise, or may not have various types of noise.
Step 2: the image detection unit response capability is calculated. To realize the difference of the responses of different detection units, the response capability of each detection unit of the original image needs to be calculated. The concrete formula is
Ni＝Di/μ (1)
Wherein N isiAs the response capability of the i-th detection unit, DiThe average value of all gray values of the ith row of the original image is obtained; μ is the mean of the entire image. The Ni value is approximately 1, and the overall response difference W of all the detection units can be expressed as:
W＝∑|1-Ni| (2)
where W represents the response difference of the overall image. The larger the value of W, the larger the difference in response.
And 3, smoothing and filtering the response capability of the image detection unit. And smoothing the pixel capability Ni of all the detection units, wherein the smoothing method is a window median filtering method. The corresponding formula is:
N′i＝median([Ni-k,Ni+k]) (3)
wherein 2k +1 is window size, N'iIs the response capability of the detection unit after spectral smoothing.
And 4, calculating the variation difference of the response capability of each detection unit after filtering, and marking the column if the variation difference is larger than a set threshold.
Calculating the difference value of the response capability of the detection unit before and after the spectrum smoothing, wherein when the difference value is greater than a certain threshold value, the detection unit is a detection unit without stripe noise; otherwise, the detection unit containing noise is considered, and the position and the difference value of the detection unit are determined. The difference value calculation formula is as follows:
Ci＝N′i-Ni (4)
wherein, CiThe difference is responded to by the ith detection unit.
Step 5, if no marked column exists, no compensation is carried out, and the image is output; if a marked detection unit is present, the stripe noise type is determined.
And judging the type of the strip noise by utilizing the respective characteristics of the periodic noise and the bad line noise. When the positions of the detection units containing the noise are continuous probe numbers, namely the probe numbers of two adjacent noises are less than 5, the noise is judged to be periodic noise. Otherwise, the noise is judged as bad line noise.
And 6, outputting the strip noise position. When the noise is judged to be periodic noise, the covered columns are more, only the serial numbers from the first detection noise position to the last detection noise position are output, namely the serial numbers of the start column and the end column, and the coverage range is the periodic noise. And when the bad line noise is judged, the position corresponding to the column with the difference value exceeding the set threshold value is the position of the bad line noise, and the serial number of the column is output.
Step 7 strip noise removal. And when the noise is judged to be periodic noise, effectively removing the stripe noise by adopting an improved moment matching algorithm. The improved moment matching algorithm formula is as follows:
Dcal_i＝DiGi (5)
wherein: DNcal_iMean value of gray values compensated for ith column, GiThe normalized gain for the ith probe is typically close to 1. The normalized gain can be expressed as:
Gi＝N′i/Ni (6)
and when the line noise is judged to be bad, removing by adopting a left-right column averaging method. The corresponding formula is:
D′i＝(Di-1+Di+1)/2 (7)
another aspect of the present invention provides a system for removing image stripe noise based on spectral filtering, which is combined with fig. 5 to implement automatic removal of image noise. The device comprises an image reading unit, a response capability calculating unit, a filtering unit, a difference value marking unit, a judging unit, a periodic noise compensating unit, a bad line noise compensating unit and an output unit;
the image reading unit reads a panchromatic or multispectral image.
The response capability calculation unit calculates the response capability of each detection unit in the image; calculating the mean value D of the gray values of each column of the imageiAnd the mean value mu of the gray value of the whole image; calculating the response capability N of the ith detection uniti＝Di/μ。
The filtering unit filters the response capability of each detection unit; the method comprises the following steps: receiving the response capability N of the ith detection unit output by the response capability calculation unitiSetting a filter window with the ith detection unit as the center, wherein the size of the filter window is 2k +1, and calculating the average value of the response capability of each detection unit in the filter window as the filtered response capability N 'of the ith detection unit'i。
The difference marking unit marks the row if the variation difference of the response capability of each detection unit after filtering is calculated and is greater than a set threshold; variation difference is Ci＝N′i-NiAnd C isiThe threshold value is set to 5 in comparison with the set threshold value.
The judging unit judges, if no marked column exists, compensation is not carried out, and the output unit directly outputs the image; if the marked detection unit exists, calculating the difference between the marked row serial numbers, if the difference between the row serial numbers is smaller than a threshold, judging that the periodic noise exists, and outputting the starting row and the ending row of each group of the marked periodic noise to a periodic noise compensation unit; if the difference value of the serial numbers of the columns is not less than the threshold, the noise of the bad line is judged, and the marked columns are output to a noise compensation unit of the bad line;
the periodic noise compensation unit compensates each column covered by the initial column and the final column by adopting a gain adjustment method; the method comprises the following steps: calculating the normalized gain G of the ith detection uniti＝N′i/Ni(ii) a Calculating the gray value mean value D after the i-th column compensationcal_i＝DiGi。
The bad line noise compensation unit compensates the marked column by adopting a left-right column averaging method; the method comprises the following steps: calculating the gray value mean value D 'of the compensated i-th column'i＝(Di-1+Di+1)/2。
The output unit outputs the image compensated by the periodic noise compensation unit and the bad line noise compensation unit.
Further, if only periodic noise or bad line noise exists, the compensated image is output by the output unit after being compensated by the corresponding unit. If the periodic noise and the bad line noise exist at the same time, the periodic noise compensation unit compensates the periodic noise of the image, then the bad line noise compensation unit compensates the bad line noise and outputs the image to the output unit, or vice versa, the bad line noise compensation unit compensates the bad line noise, then the periodic noise compensation unit compensates the periodic noise of the image and outputs the image to the output unit.
Furthermore, if there may exist multiple groups of periodic noise and bad line noise in the image, the compensation of each group of noise of the same type of noise is completed, and then the compensation of each group of noise of another type of noise is performed.
Examples
To effectively illustrate the stripe noise of the image, firstly, an image with stripe noise is simulated by an analog simulation method, as shown in fig. 2. Its noise includes two categories, bad line noise and periodic stripe noise. The positions of the bad line noise are 173 columns and 450 columns, the number of the periodic stripe noises is two, the first periodic stripe noise ranges from 240 columns to 290 columns respectively, and the second periodic stripe noise ranges from 330 columns to 380 columns.
Firstly, reading a remote sensing image, and calculating a detection unit response capability evaluation index Ni of the image according to a formula 1 to obtain a response capability evaluation index curve of each detection element.
And then, performing smooth filtering processing on the probe element response capability evaluation index curve by adopting a median filtering technology to obtain a smoothed probe element response capability evaluation index curve graph. As shown in fig. 3.
And calculating the difference of the detection response capabilities before and after smoothing. If the absolute value of the difference is greater than a given threshold (tentatively 0.05 here), the probe unit is considered to contain noise, and the corresponding probe number is recorded.
And counting all the probe numbers containing noise, and if the probe numbers are a plurality of continuous probes, considering the probe as periodic stripe noise. The judgment basis of the continuous detecting element is that the number interval of two adjacent detecting elements containing noise is less than 5. Through calculation, two periodic noises can be obtained, wherein the starting and ending probe numbers of the first periodic stripe noise are 240 and 290 respectively, and the starting and ending probe numbers of the second periodic stripe noise are 330 and 380 respectively. The result is the same as the set value of the analog noise image.
It is calculated that there are two bad line noises, corresponding position distributions 173 and 450, and the same setting values as the analog noise image.
Compensating the periodic noise by adopting a gain adjustment method; and for the bad line noise, the effective removal of the bad line noise is realized by using an averaging method of gray values at the left side and the right side of the probe element. An image subjected to periodic noise and bad line noise removal is shown in fig. 4.
In summary, the present invention relates to a method and a system for removing noise in image bands based on spectral filtering, which reads panchromatic or multispectral images; identifying periodic noise and bad line noise in the image; compensating the periodic noise by adopting a gain adjustment method; and compensating the bad line noise by adopting a left-right column averaging method. The invention can realize the determination of the type of the strip noise and the rapid determination of the position of the strip noise by defining and calculating indexes such as the response capability of the detection unit, the response difference of the detection unit and the like of the image, the whole processing process is automatically realized by using a computer, the manual intervention is not needed, the processing speed is high, the automation degree is high, and the automatic batch processing of mass images can be realized. The method is suitable for satellite images with irregular stripe noise, and can be used for quickly identifying and removing the stripe noise positions and types of different types of remote sensing images such as panchromatic images, multispectrals images, hyperspectral images and the like.
It is to be understood that the above-described embodiments of the present invention are merely illustrative of or explaining the principles of the invention and are not to be construed as limiting the invention. Therefore, any modification, equivalent replacement, improvement and the like made without departing from the spirit and scope of the present invention should be included in the protection scope of the present invention. Further, it is intended that the appended claims cover all such variations and modifications as fall within the scope and boundaries of the appended claims or the equivalents of such scope and boundaries.
Claims (10)
1. An image strip noise removing method based on spectral filtering is characterized by comprising the following steps:
reading a panchromatic or multispectral image;
calculating the response capability of each detection unit in the image;
filtering the response capability of each detection unit;
calculating the variation difference value of the response capability of each detection unit after filtering, and marking the column if the variation difference value is larger than a set threshold value;
if no marked column exists, no compensation is carried out, and the image is output; if a marked detection unit is present:
calculating the difference between the marked row serial numbers, if the difference between the row serial numbers is smaller than a threshold, determining that the row serial numbers are periodic noise, and compensating the periodic noise of the row by adopting a gain adjustment method; otherwise, judging the bad line noise, and compensating the bad line noise of the row by adopting a left-right row averaging method.
2. The method according to claim 1, wherein calculating the response capability of each detection unit in the image comprises: calculating the mean value D of the gray values of each column of the imageiAnd the mean value mu of the gray value of the whole image; calculating the response capability N of the ith detection uniti＝Di/μ。
3. The method of claim 2, wherein the filtering comprises: setting a filter window by taking the ith detection unit as a center, wherein the size of the filter window is 2k +1, and calculating the average value of the response capability of each detection unit in the filter window as the filtered response capability N 'of the ith detection unit'i. Further, the variation difference of the ith detection unit is Ci＝N′i-NiThe set threshold is 5.
4. The spectral filtering based image strip noise removing method according to claim 3, wherein the compensating the periodic noise of the column by using a gain adjustment method comprises:
calculating the normalized gain G of the ith detection uniti＝N′i/Ni；
Calculating the gray value mean value D after the i-th column compensationcal_i＝DiGi；
Adopting a left-right column average method to compensate the bad line noise of the column, comprising the following steps:
calculating the gray value mean value D 'of the compensated i-th column'i＝(Di-1+Di+1)/2。
5. The spectral filtering based image strip noise removal method according to claim 3 or 4, characterized in that only the start column and the end column are marked for periodic noise; the column where the noise is located is marked for bad line noise.
6. An image strip noise removing system based on spectral filtering is characterized by comprising an image reading unit, a response capability calculating unit, a filtering unit, a difference marking unit, a judging unit, a periodic noise compensating unit, a bad line noise compensating unit and an output unit;
the image reading unit reads a panchromatic or multispectral image;
the response capability calculation unit calculates the response capability of each detection unit in the image;
the filtering unit filters the response capability of each detection unit;
the difference marking unit marks the row if the variation difference of the response capability of each detection unit after filtering is calculated and is greater than a set threshold;
the judging unit judges, if no marked column exists, compensation is not carried out, and the output unit directly outputs the image; if the marked detection unit exists, calculating the difference between the marked row serial numbers, if the difference between the row serial numbers is smaller than a threshold, judging that the periodic noise exists, and outputting the starting row and the ending row of each group of the marked periodic noise to a periodic noise compensation unit; if the difference value of the serial numbers of the columns is not less than the threshold, the noise of the bad line is judged, and the marked columns are output to a noise compensation unit of the bad line;
the periodic noise compensation unit compensates each column covered by the initial column and the final column by adopting a gain adjustment method;
the bad line noise compensation unit compensates the marked column by adopting a left-right column averaging method;
the output unit outputs the image compensated by the periodic noise compensation unit and the bad line noise compensation unit.
7. The system according to claim 6, wherein the response capability calculating unit calculates the response capability of each detecting unit in the image comprises: calculating the mean value D of the gray values of each column of the imageiAnd the mean value mu of the gray value of the whole image; calculating the response capability N of the ith detection uniti＝Di/μ。
8. The system according to claim 7, wherein the filtering unit filters the response capability of each detection unit, and comprises: receiving the response of the ith detection unit output by the response capability calculation unitCapability NiSetting a filter window with the ith detection unit as the center, wherein the size of the filter window is 2k +1, and calculating the average value of the response capability of each detection unit in the filter window as the filtered response capability N 'of the ith detection unit'i。
Further, the difference marking unit calculates the variation difference of the ith detection unit as Ci＝N′i-NiAnd C isiCompared with a set threshold, the set threshold is 5.
9. The system according to claim 8, wherein the periodic noise compensation unit compensates for the covered columns of the start column and the end column by a gain adjustment method, comprising: calculating the normalized gain G of the ith detection uniti＝N′i/Ni(ii) a Calculating the gray value mean value D after the i-th column compensationcal_i＝DiGi。
10. The system according to claim 8 or 9, wherein the bad line noise compensation unit compensates the marked columns by left-right column averaging, and comprises: calculating the gray value mean value D 'of the compensated i-th column'i＝(Di-1+Di+1)/2。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010850347.0A CN112017127A (en) | 2020-08-21 | 2020-08-21 | Image strip noise removing method and system based on spectral filtering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010850347.0A CN112017127A (en) | 2020-08-21 | 2020-08-21 | Image strip noise removing method and system based on spectral filtering |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112017127A true CN112017127A (en) | 2020-12-01 |
Family
ID=73505536
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010850347.0A Pending CN112017127A (en) | 2020-08-21 | 2020-08-21 | Image strip noise removing method and system based on spectral filtering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112017127A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020913A (en) * | 2012-12-18 | 2013-04-03 | 武汉大学 | Remote-sensing image stripe noise removing method based on segmented correction |
| CN103761718A (en) * | 2014-02-12 | 2014-04-30 | 北京空间机电研究所 | Satellite remote sensing image region stripe noise suppression device and method thereof |
| CN106504208A (en) * | 2016-10-27 | 2017-03-15 | 西京学院 | Based on orderly minima and the high-spectrum image width destriping method of wavelet filtering |
-
2020
- 2020-08-21 CN CN202010850347.0A patent/CN112017127A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103020913A (en) * | 2012-12-18 | 2013-04-03 | 武汉大学 | Remote-sensing image stripe noise removing method based on segmented correction |
| CN103761718A (en) * | 2014-02-12 | 2014-04-30 | 北京空间机电研究所 | Satellite remote sensing image region stripe noise suppression device and method thereof |
| CN106504208A (en) * | 2016-10-27 | 2017-03-15 | 西京学院 | Based on orderly minima and the high-spectrum image width destriping method of wavelet filtering |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109697476B (en) | X-ray photon counting detector consistency calibration method based on deep learning | |
| CN115619793B (en) | Power adapter appearance quality detection method based on computer vision | |
| CN116012346B (en) | Power transmission line monitoring method and system | |
| US20090034827A1 (en) | Inspection device, inspection method, method of manufacturing color filter, and computer-readable storage medium containing inspection device control program | |
| CN103400357B (en) | A kind of method removing image salt-pepper noise | |
| US6912322B2 (en) | Adaptive process for removing streaks in multi-band digital images | |
| CN107727231B (en) | Method for correcting stripe noise of spectral imager | |
| JP2013508714A (en) | How to analyze the quality of glazing units | |
| CN115272303B (en) | Textile fabric defect degree evaluation method, device and system based on Gaussian blur | |
| US7796153B1 (en) | Equalization system and method for an imaging sensor | |
| CN112435181A (en) | Method for filtering vertical stripe noise of uncooled infrared video image | |
| CN116245880A (en) | Electric vehicle charging pile fire risk detection method based on infrared identification | |
| CN115797473A (en) | Concrete forming evaluation method for civil engineering | |
| CN111353968B (en) | Infrared image quality evaluation method based on blind pixel detection and analysis | |
| CN112053299B (en) | Image processing method based on colloidal gold immunochromatography | |
| US6987577B2 (en) | Providing a partial column defect map for a full frame image sensor | |
| CN112017127A (en) | Image strip noise removing method and system based on spectral filtering | |
| CN111915682A (en) | Real-time self-adjusting hyperspectral camera non-uniformity correction method | |
| JP2006067423A (en) | Image quality quantitative evaluation method and its device | |
| CN115760749A (en) | Millimeter wave radiation image fire detection and identification method and system | |
| WO2012063265A2 (en) | Method and apparatus for detecting the bad pixels in sensor array and concealing the error | |
| DK2691920T3 (en) | IDENTIFICATION AND CORRECTION OF ANOMALIES IN AN OPTICAL IMAGE | |
| CN113436096A (en) | Push-broom hyperspectral imaging strip noise elimination method based on pixel calibration | |
| Fischer et al. | Median spectral-spatial bad pixel identification and replacement for hyperspectral SWIR sensors | |
| CN115311164A (en) | Synthetic beam effect eliminating method for sparse interference array observation imaging |
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 |