CN116823764A - Infrared image blind pixel detection method based on sliding window - Google Patents

Infrared image blind pixel detection method based on sliding window Download PDF

Info

Publication number
CN116823764A
CN116823764A CN202310758738.3A CN202310758738A CN116823764A CN 116823764 A CN116823764 A CN 116823764A CN 202310758738 A CN202310758738 A CN 202310758738A CN 116823764 A CN116823764 A CN 116823764A
Authority
CN
China
Prior art keywords
blind
pixel
candidate
pixels
point
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
Application number
CN202310758738.3A
Other languages
Chinese (zh)
Inventor
刘佳
贾格
戴国民
魏宇星
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Institute of Optics and Electronics of CAS
Original Assignee
Institute of Optics and Electronics of CAS
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Institute of Optics and Electronics of CAS filed Critical Institute of Optics and Electronics of CAS
Priority to CN202310758738.3A priority Critical patent/CN116823764A/en
Publication of CN116823764A publication Critical patent/CN116823764A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/70—Determining position or orientation of objects or cameras
    • G06T7/77—Determining position or orientation of objects or cameras using statistical methods
    • 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
    • G06T2207/10048—Infrared image

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Probability & Statistics with Applications (AREA)
  • Transforming Light Signals Into Electric Signals (AREA)

Abstract

The invention discloses an infrared image blind pixel detection method based on a sliding window, which comprises the steps of firstly taking a current point as a center, selecting a window with the size of 3x3, and judging whether the current point is a candidate blind pixel; and judging whether the blind pixel is a real blind pixel according to the difference value between the candidate blind pixel and 8 surrounding pixel points and the difference value between the 8 surrounding pixel points, and if the candidate blind pixel point is not determined to be the blind pixel, continuously judging whether the point is the blind pixel according to the maximum value or the minimum value of the 8 surrounding pixel points. The invention is suitable for blind pixel detection of infrared images in various scenes, can detect fixed blind pixels, and has good detection effect on random blind pixels.

Description

Infrared image blind pixel detection method based on sliding window
Technical Field
The invention relates to an infrared image blind pixel detection method based on a sliding window, and relates to the field of digital image processing.
Background
The infrared detector is an important device in an infrared imaging system, and is influenced by various factors such as external environment, infrared sensitive elements, circuit structures, semiconductor characteristics and the like, blind pixels are commonly arranged in an infrared image, and the blind pixels can lead the image to have bright spots or dark spots. The existence of the blind pixels seriously affects the imaging quality of an infrared system, affects the subsequent image analysis and processing work, and easily causes the condition of missing the detection target when the weak target is detected.
The method for detecting the blind pixels of the infrared image based on the sliding window is provided in order to rapidly and accurately realize the blind pixel detection.
Disclosure of Invention
The technical problems to be solved by the invention are as follows: the infrared image blind pixel detection method overcomes the defects of the prior art and provides a sliding window infrared image blind pixel detection method.
The technical scheme adopted for solving the technical problems is as follows:
an infrared image blind pixel detection method of a sliding window, comprising the following steps:
selecting a window with the size of 3x3 by taking the current pixel point as the center, and judging whether the current pixel point is a candidate blind pixel;
and judging whether the blind pixel is a real blind pixel according to the difference value between the candidate blind pixel and 8 surrounding pixel points and the difference value between the 8 surrounding pixel points, and if the candidate blind pixel point is not determined to be the blind pixel, continuously judging whether the point is the blind pixel according to the maximum value or the minimum value of the 8 surrounding pixel points.
Further, for the input image, a window with a size of 3x3 is selected with each pixel point as a center, and if the pixel point is a maximum value or a minimum value in the window, the pixel point is classified as a candidate blind pixel.
Further, for the candidate blind pixels (r, c) (row coordinates of the candidate blind pixels of r, c respectively), calculating absolute values of differences between the candidate blind pixels (r, c) and 8 pixel values around the window, taking out a maximum value t1, then taking out 8 pixel points around the candidate blind pixels, calculating absolute values of differences between two adjacent pixel points, taking out a maximum value t2, if:
t1＞3*t2+n
wherein n is a constant, 5 < n < 15, then the blind candidate (r, c) is determined to be a blind candidate, the blind candidate (r, c) is replaced by the median point in the window, if:
t1＜t2+n
then the candidate blind pixel (r, c) is determined to be a non-blind pixel if:
t2+n≤t1≤3*t2+n
then the candidate blind pixel (r, c) is still a candidate blind pixel.
Further, if the blind candidate (r, c) is still the blind candidate, a maximum point (r) is found from 8 pixel points around the blind candidate (r, c) max ,c max )(r max ,c max Row coordinates of the respective maximum points), with a maximum point (r max ,c max ) Taking a 3x3 window as the center, calculating the maximum point (r max ,c max ) Taking the maximum value m1 from the absolute value of the difference between the adjacent 7 pixels (eliminating the candidate blind pixels (r, c)) around, then calculating the absolute value of the difference between the adjacent two points of the 7 pixels around, taking the maximum value m2, and if:
m1＞3*m2+n
the maximum point (r max ,c max ) And candidate blind pixels (r, c) are determined as blind pixels simultaneously, and the maximum points (r max ,c max ) And candidate blind pixels (r, c) as centers, taking a 3x3 window, and replacing the maximum point (r) with the median value in the window max ,c max ) And candidate blind pixels (r, c), otherwise, the candidate blind pixels (r, c) are still candidate blind pixels.
Further, if the blind candidate (r, c) is still the blind candidate, a minimum value point (r) is found from 8 pixel points around the blind candidate (r, c) min ,c min )(r min ,c min Row coordinates of the respective maximum points), with minimum points (r min ,c min ) Taking a 3x3 window as the center, calculating a minimum point (r min ,c min ) Taking the maximum value f1 from the absolute value of the difference between the adjacent 7 pixels (eliminating the candidate blind pixels (r, c)) around, then calculating the absolute value of the difference between the adjacent two points of the 7 pixels around, taking the maximum value f2, if:
f1＞3*f2+n
then the minimum point (r min ,c min ) And candidate blind pixels (r, c) are determined as blind pixels simultaneously, and the minimum blind pixels are respectively usedValue point (r) min ,c min ) And candidate blind pixels (r, c) as centers, taking a 3x3 window, and replacing (r) with the median value in the window min ,c min ) And candidate blind pixels (r, c), otherwise, the candidate blind pixels (r, c) are determined to be non-blind pixels.
The beneficial effects of the invention are as follows:
1. the invention is suitable for blind pixel detection of infrared images in various scenes, can detect fixed blind pixels, and has good detection effect on random blind pixels;
2. the invention has better detection effect on single-point and continuous two-point blind pixels;
3. the invention has good effect on detecting the blind pixels of the infrared image with the weak and small targets, can keep the weak and small targets while detecting the blind pixels, and avoids the weak and small targets from being detected as the blind pixels;
4. the invention has the advantages of simple model, low operation complexity and high operation speed, and is suitable for various blind pixel detection systems with high real-time requirements.
Drawings
FIG. 1 is a flow chart of an infrared image blind pixel detection method based on a sliding window;
FIG. 2 is a 3x3 size window centered on a blind candidate, p 1 ～p 8 Representing 8 pixel points around the candidate blind pixel;
FIG. 3 is a 3x3 size window centered on a blind candidate and a maximum point around the blind candidate, p 1 ～p 7 7 pixel points around the maximum point are represented (blind candidate pixels are removed);
FIG. 4 is a 3x3 size window centered on a blind candidate and a minimum point around the blind candidate, p 1 ～p 7 7 pixel points around the minimum value point are represented (candidate blind pixels are removed);
fig. 5 shows the blind pixel detection and correction result of the infrared weak small target image according to the present invention, wherein (a) is the original image and (b) is the blind pixel detection and correction result.
Detailed Description
The invention is further described below with reference to the drawings and detailed description.
As shown in fig. 1, the method for detecting the blind pixels of the infrared image based on the sliding window specifically comprises the following steps:
first, for an input image, a window with a size of 3x3 is selected with each pixel point as a center, and if the pixel point is a maximum value or a minimum value in the window, the pixel point is classified as a candidate blind pixel.
Second, for the blind candidate pixels (r, c) (the row coordinates of the blind candidate pixels are respectively calculated, the blind candidate pixels (r, c) and 8 pixel values p around the window are calculated 1 ～p 8 As shown in fig. 2, taking the maximum value t1, then taking out 8 pixel points around the blind pixel candidate, calculating the absolute value of the difference between two adjacent pixel points, taking the maximum value t2, if:
t1＞3*t2+n
wherein n is a constant, 5 < n < 15, then the blind candidate (r, c) is determined to be a blind candidate, the blind candidate (r, c) is replaced by the median point in the window, if:
t1＜t2+n
then the candidate blind pixel (r, c) is determined to be a non-blind pixel if:
t2+n≤t1≤3*t2+n
then the candidate blind pixel (r, c) is still a candidate blind pixel.
Third, if the blind candidate (r, c) is still the blind candidate after the second step is performed, searching the maximum value point (r) from 8 pixel points around (r, c) max ,c max )(r max ,c max Row coordinates of the respective maximum points), with a maximum point (r max ,c max ) Taking a 3x3 window as the center, calculating the maximum point (r max ,c max ) And 7 pixel points p around 1 ～p 7 (reject candidate blind pixels (r, c)) as shown in fig. 3, take the maximum value m1, then calculate the absolute value of the difference between two adjacent points of the surrounding 7 pixels, take the maximum value m2, if:
m1＞3*m2+n
the maximum point (r max ,c max ) And candidate blind pixels (r, c) are determined as blind pixels simultaneously, and the maximum points (r max ,c max ) And candidate blind pixels (r, c) as centers, taking a 3x3 window, and replacing the maximum point (r) with the median value in the window max ,c max ) And candidate blind pixels (r, c), otherwise, the candidate blind pixels (r, c) are still candidate blind pixels.
Fourth, if the blind candidate (r, c) is still the blind candidate after the third step is performed, searching the minimum value point (r from 8 pixel points around the blind candidate (r, c) min ,c min )(r min ,c min Row coordinates of the respective maximum points), with minimum points (r min ,c min ) Taking a 3x3 window as the center, calculating a minimum point (r min ,c min ) And 7 pixel points p around 1 ～p 7 (reject candidate blind pixels (r, c)) taking the maximum value f1 as shown in fig. 4, then calculating the absolute value of the difference between two adjacent points of the surrounding 7 pixel points, taking the maximum value f2, if:
f1＞3*f2+n
then the minimum point (r min ,c min ) And candidate blind pixels (r, c) are determined as blind pixels simultaneously, and the minimum value points (r min ,c min ) And candidate blind pixels (r, c) as centers, taking a 3x3 window, and replacing a minimum value point (r) with a median value in the window min ,c min ) And candidate blind pixels (r, c), otherwise, the candidate blind pixels (r, c) are determined to be non-blind pixels.
In fig. 5, (a) is an original infrared image, which contains a small target and a plurality of blind pixels, and (b) is an image after blind pixel detection and correction.
Parts of the invention not described in detail are well known in the art. The above examples are merely illustrative of preferred embodiments of the invention, which are not exhaustive of all details, nor are they intended to limit the invention to the particular embodiments disclosed. Various modifications and improvements of the technical scheme of the present invention will fall within the protection scope of the present invention as defined in the claims without departing from the design spirit of the present invention.

Claims (5)

1. The method for detecting the blind pixels of the infrared image based on the sliding window is characterized by comprising the following steps:
selecting a window with the size of 3x3 by taking the current pixel point as the center, and judging whether the current pixel point is a candidate blind pixel;
and judging whether the blind pixel is the blind pixel according to the difference value between the candidate blind pixel and 8 surrounding pixel points and the difference value between the 8 surrounding pixel points, and if the candidate blind pixel is not determined to be the blind pixel, continuously judging whether the current pixel is the blind pixel according to the maximum value or the minimum value of the 8 surrounding pixel points.
2. The method for detecting blind pixels in an infrared image based on a sliding window according to claim 1, wherein for an input image, a window with a size of 3x3 is selected with each pixel point as a center, and if the pixel point is a maximum value or a minimum value in the window, the pixel point is classified as a candidate blind pixel.
3. The method for detecting blind pixels of infrared image based on sliding window according to claim 2, wherein, for the blind pixel candidates (r, c), the row coordinates of the blind pixel candidates are calculated, the absolute value of the difference between the blind pixel candidates (r, c) and 8 pixels around the inside of the window is taken out, the maximum value t1 is taken out, then 8 pixels around the blind pixel candidates are taken out, the absolute value of the difference is calculated between two adjacent pixels, the maximum value t2 is taken out, if:
t1＞3*t2+n
wherein n is a constant, 5 < n < 15, then the blind candidate (r, c) is determined to be a blind candidate, the blind candidate (r, c) is replaced by the median point in the window, if:
t1＜t2+n
then the candidate blind pixel (r, c) is determined to be a non-blind pixel if:
t2+n≤t1≤3*t2+n
then the candidate blind pixel (r, c) is still a candidate blind pixel.
4. A sliding window based infrared image blind pixel detection method according to claim 3, wherein if the blind pixel candidate (r, c) is still a blind pixel candidateThe element searches the maximum value point (r) from 8 pixel points around the candidate blind element (r, c) max ,c max ) Wherein r is max ,c max The row and column coordinates of the maximum points are respectively represented by the maximum point (r max ,c max ) Taking a 3x3 window as the center, calculating the maximum point (r max ,c max ) Taking the maximum value m1 from the absolute values of the differences of the surrounding 7 pixel points after the candidate blind pixels (r, c) are removed, then calculating the absolute values of the differences of the adjacent two points of the surrounding 7 pixel points, taking the maximum value m2, and if:
m1＞3*m2+n
the maximum point (r max ,c max ) And candidate blind pixels (r, c) are determined as blind pixels simultaneously, and the maximum points (r max ,c max ) And candidate blind pixels (r, c) as centers, taking a 3x3 window, and replacing the maximum point (r) with the median value in the window max ,c max ) And candidate blind pixels (r, c), otherwise, the candidate blind pixels (r, c) are still candidate blind pixels.
5. The method for detecting blind pixels of infrared image based on sliding window according to claim 4, wherein if the blind pixel candidate (r, c) is still the blind pixel candidate, the minimum value point (r min ,c min ) Wherein r is min ,c min The row and column coordinates of the maximum points are respectively calculated as minimum points (r min ,c min ) Taking a 3x3 window as the center, calculating a minimum point (r min ,c min ) Taking the maximum value f1 from the absolute values of the differences of the surrounding 7 pixel points after the candidate blind pixels (r, c) are removed, then calculating the absolute values of the differences of the adjacent two points of the surrounding 7 pixel points, taking the maximum value f2, and if:
f1＞3*f2+n
then the minimum point (r min ,c min ) And candidate blind pixels (r, c) are determined as blind pixels simultaneously, and the minimum value points (r min ,c min ) And candidate blind pixels (r, c) as centers, taking a 3x3 window, and replacing a minimum value point (r) with a median value in the window min ,c min ) And candidate blind pixels (r, c), otherwise, the candidate blind pixels (r, c) are determined to be non-blind pixels.
CN202310758738.3A 2023-06-26 2023-06-26 Infrared image blind pixel detection method based on sliding window Pending CN116823764A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310758738.3A CN116823764A (en) 2023-06-26 2023-06-26 Infrared image blind pixel detection method based on sliding window

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310758738.3A CN116823764A (en) 2023-06-26 2023-06-26 Infrared image blind pixel detection method based on sliding window

Publications (1)

Publication Number Publication Date
CN116823764A true CN116823764A (en) 2023-09-29

Family

ID=88121533

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310758738.3A Pending CN116823764A (en) 2023-06-26 2023-06-26 Infrared image blind pixel detection method based on sliding window

Country Status (1)

Country Link
CN (1) CN116823764A (en)

Similar Documents

Publication Publication Date Title
CN114937055A (en) Image self-adaptive segmentation method and system based on artificial intelligence
US20090040343A1 (en) Methods and apparatuses for defective pixel detection and correction
CN111080582B (en) Method for detecting defects of inner and outer surfaces of workpiece
US7773783B2 (en) Face detecting method and photography apparatus using the method
JP2016048815A (en) Image processor, image processing method and image processing system
Hwang et al. Difference-based image noise modeling using skellam distribution
CN112419261B (en) Visual acquisition method and device with abnormal point removing function
CN103093458A (en) Detecting method and detecting device for key frame
CN111612773B (en) Thermal infrared imager and real-time automatic blind pixel detection processing method
CN110728668A (en) Airspace high-pass filter for maintaining small target form
CN113971681A (en) Edge detection method for belt conveyor in complex environment
JP2005172559A (en) Method and device for detecting line defect on panel
CN110853077B (en) Self-adaptive infrared dynamic frame feature extraction method based on morphological change estimation
CN113077416A (en) Welding spot welding defect detection method and system based on image processing
CA2704037A1 (en) Method for detecting a target
CN114170153A (en) Wafer defect detection method and device, electronic equipment and storage medium
CN117456371A (en) Group string hot spot detection method, device, equipment and medium
CN110887563B (en) Hyperspectral area array detector bad element detection method
CN116823764A (en) Infrared image blind pixel detection method based on sliding window
US20060126136A1 (en) Image inspecting apparatus, image inspecting method, control program and computer-readable storage medium
CN108322723B (en) Color distortion compensation method and device and television
JP2009017158A (en) Camera inspection device
RU2446471C1 (en) Method for detecting moving objects and determining parameters thereof
US20120218447A1 (en) Image pickup apparatus
CN107346533A (en) A kind of blind-pixel elimination method of view-based access control model characteristic

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