Search

CN-122023167-A - Method, device, equipment and medium for multipoint correction of infrared image

CN122023167ACN 122023167 ACN122023167 ACN 122023167ACN-122023167-A

Abstract

The invention provides a method, a device, equipment and a medium for multipoint correction of an infrared image, which comprise the steps of collecting infrared images with the temperature of more than or equal to 10 frames in a dark field, low illumination, medium illumination and high illumination in real time, carrying out multi-frame average processing on each frame of images to inhibit time domain noise, generating an image to be analyzed, extracting pixel characteristic parameters of the image to be analyzed, generating blind pixel detection parameters based on the response rate, noise voltage, noise equivalent temperature difference and linearity characteristics of each pixel, marking blind pixel elements and normal pixels of the image to be analyzed according to a constructed blind pixel table based on the blind pixel detection parameters, and carrying out average interpolation filling on the blind pixel elements by adopting self-adaptive gain and offset parameters of the normal pixels in a 3×3 neighborhood. The invention combines multi-frame data, multi-criterion blind pixel detection and multi-point temperature correction methods to realize wide-temperature-range and high-precision infrared image correction, and simultaneously combines noise suppression and pixel nonlinear compensation.

Inventors

  • SHI KE
  • YUAN LIANG
  • FANG ZHENGJUN

Assignees

  • 安徽光智科技有限公司

Dates

Publication Date
20260512
Application Date
20251224

Claims (10)

  1. 1. The method for correcting the infrared image in a multipoint manner is characterized by comprising the following steps of: Step S1, acquiring infrared images of more than or equal to 10 frames in four temperature scenes of dark field, low illumination, medium illumination and high illumination in real time, and carrying out multi-frame average processing on each frame of images to inhibit time domain noise so as to generate an image to be analyzed; S2, extracting pixel characteristic parameters of the image to be analyzed, and generating blind pixel detection parameters based on the response rate, noise voltage, noise equivalent temperature difference and linearity characteristics of each pixel; Step S3, marking out a blind pixel and a normal pixel of the image to be analyzed according to the constructed blind pixel table based on the blind pixel detection parameters; And S4, for the blind pixel, carrying out average interpolation filling by adopting the adaptive gain and offset parameters of the normal pixels in the 3 multiplied by 3 neighborhood to realize correction.
  2. 2. The method of multipoint infrared image correction according to claim 1, wherein the process of performing multi-frame averaging processing on each frame of image is implemented using the following formula: Wherein F is the number of frames, For the f frame pixel Is set in the circuit (a) and the signal voltage of the circuit (b).
  3. 3. The method of multipoint correction of an infrared image according to claim 1, wherein the calculation formulas of the response rate, the noise voltage, the noise equivalent temperature difference, and the linearity characteristics generated in the step S2 are expressed as: In the above-mentioned method, the step of, Representing picture elements at different temperatures Is used for the response rate of the (a), 、 Pixels with temperature T in high-illumination scene and temperature T0 in low-illumination scene respectively A signal voltage; Representing a picture element in a dark field Is set to be a noise voltage of a (c), Representing a picture element in a dark field Is used for the response voltage of the (a), Representing the f-th frame pixel at the low illumination scene temperature T0 Is set to the signal voltage of (1); Representing the noise equivalent temperature difference; A non-linear deviation representing the characteristic of the linearity, Representing picture elements The original signal voltage at temperature T in a high light scene, 、 Representing picture elements Response curve fitting coefficients of (a).
  4. 4. The method for multipoint correction of an infrared image according to claim 3, wherein the step of constructing a blind pixel table based on blind pixel detection parameters specifically comprises the steps of fusing a spatial response criterion, a response rate criterion, a time domain noise criterion, a linearity criterion and a G parameter criterion, generating the blind pixel table through logical OR operation, wherein when the blind pixel of the image to be analyzed is marked through the blind pixel table, any performance parameter of the spatial response criterion, the response rate criterion, the time domain noise criterion, the linearity criterion and the G parameter criterion is abnormal, and outputting a detection result of the blind pixel.
  5. 5. The method for multipoint correction of an infrared image according to claim 4, wherein the spatial response criterion is specifically that a first threshold interval is generated according to a global average value and a global standard deviation value of pixel response values in dark field, medium illumination and high illumination different temperature scenes, and if the pixel response value is greater than a maximum value of the first threshold interval or less than a minimum value of the first threshold interval, a detection result of a blind pixel is output; the response rate criterion is specifically that a second threshold interval is generated according to the average value and the standard difference value of pixel response value differences under different temperature scenes of high illumination and dark fields, and if the pixel response value difference value is larger than the maximum value of the second threshold interval or smaller than the minimum value of the second threshold interval, a detection result of a blind pixel is output; The time domain noise criterion is specifically that a third threshold interval is generated according to the mean value and the standard deviation value of noise standard deviation values in dark fields, medium illumination and high illumination under different temperature scenes, and if the time domain noise value is larger than the maximum value of the third threshold interval or smaller than the minimum value of the third threshold interval, a detection result of a blind pixel is output; the linearity criterion is specifically that a nonlinear deviation value is generated according to a slope and an intercept obtained by fitting pixel response values under different temperature scenes of high illumination and dark fields, and if the nonlinear deviation value is larger than a fourth threshold value, a blind pixel detection result is output; The G parameter criterion is specifically that if the correction gain parameter is greater than the maximum value of the fifth threshold interval or less than the minimum value of the fifth threshold interval, a detection result of the blind pixel is output.
  6. 6. The method of multipoint correction of an infrared image according to claim 5, wherein for blind pixels, performing an average interpolation padding correction using adaptive gain and offset parameters of normal pixels in a 3 x 3 neighborhood comprises using the following formula: In the above-mentioned method, the step of, Representing blind pixel At the temperature of The lower corrected signal voltage, N represents the number of normal pixels in the 3×3 neighborhood, Representing blind pixel Coordinates of normal pixels in a3 x 3 neighborhood, Representing neighborhood normal pixels Corrected signal voltage at temperature T.
  7. 7. The method of infrared image multipoint correction according to claim 6, further comprising correcting a correction response curve of pixel fitting based on nonlinear deviation for normal pixels, expressed as: Wherein, the Representing picture elements An adaptive gain parameter of (a), Representing picture elements Is used to determine the adaptive offset parameter of the (c), Representing picture elements At the temperature of The corrected signal voltage; Representing picture elements At the temperature of The original signal voltage below.
  8. 8. An apparatus for multi-point correction of an infrared image to perform the method of any of claims 1-7, the apparatus comprising: The image acquisition module is used for acquiring infrared images of more than or equal to 10 frames in four temperature scenes of dark field, low illumination, medium illumination and high illumination in real time, and carrying out multi-frame average processing on each frame of image to inhibit time domain noise so as to generate an image to be analyzed; The feature extraction module is used for extracting pixel feature parameters of the image to be analyzed and generating blind pixel detection parameters based on the response rate, noise voltage, noise equivalent temperature difference and linearity features of each pixel; the blind pixel marking module is used for marking out the blind pixel and the normal pixel of the image to be analyzed according to the constructed blind pixel table based on the blind pixel detection parameters; And the blind pixel correction module is used for carrying out average interpolation filling on the blind pixel by adopting the adaptive gain and offset parameters of the normal pixels in the 3 multiplied by 3 neighborhood so as to realize correction.
  9. 9. An electronic device comprising a memory and a processor, the memory having stored thereon a computer program, characterized in that the processor, when executing the computer program, implements the method according to any of claims 1-7.
  10. 10. A computer readable storage medium, on which a computer program is stored, which computer program, when being executed by a processor, implements the method according to any of claims 1-7.

Description

Method, device, equipment and medium for multipoint correction of infrared image Technical Field The present invention relates to the field of image processing technologies, and in particular, to a method, an apparatus, a device, and a medium for multi-point correction of an infrared image. Background The infrared detector has the problems of non-uniformity of pixel response, blind pixels (dead pixels and overheated pixels), noise interference and the like due to the limitation of a manufacturing process, so that the image quality is reduced. In the prior art, the two-point correction method, the single-point correction method, the blind pixel detection, the multi-frame average processing and other modes are adopted to carry out the multi-point correction of the image, however, the following technical problems exist: (1) The adaptability is poor, the two-point correction only covers a limited temperature point, and response drift in a wide temperature range scene cannot be dealt with; (2) The blind pixel detection precision is low, namely, a single criterion (such as only through response rate) is difficult to distinguish the real blind pixel from noise interference, so that the corrected image still has defects; (3) The noise suppression is insufficient, the time domain noise suppression is limited by simple multi-frame average, and the effect is poor especially in a low-illumination scene; (4) The pixel nonlinearity is not considered, namely, the traditional method assumes that the pixel response is linear, and the actual nonlinear response leads to correction deviation to influence the target detection precision. Disclosure of Invention In order to solve the technical problems, the invention provides a method, a device, equipment and a medium for infrared image multipoint correction, the invention combines multi-frame data, multi-criterion blind pixel detection and multi-point temperature correction methods to realize wide-temperature-range and high-precision infrared image correction, and simultaneously combines noise suppression and pixel nonlinear compensation. In a first aspect, the present invention provides a method for multipoint correction of an infrared image, specifically including the following steps: Step S1, acquiring infrared images of more than or equal to 10 frames in four temperature scenes of dark field, low illumination, medium illumination and high illumination in real time, and carrying out multi-frame average processing on each frame of images to inhibit time domain noise so as to generate an image to be analyzed; S2, extracting pixel characteristic parameters of the image to be analyzed, and generating blind pixel detection parameters based on the response rate, noise voltage, noise equivalent temperature difference and linearity characteristics of each pixel; Step S3, marking out a blind pixel and a normal pixel of the image to be analyzed according to the constructed blind pixel table based on the blind pixel detection parameters; And S4, for the blind pixel, carrying out average interpolation filling by adopting the adaptive gain and offset parameters of the normal pixels in the 3 multiplied by 3 neighborhood to realize correction. Further, the process of carrying out multi-frame average processing on each frame of image is realized by adopting the following formula: Wherein F is the number of frames, For the f frame pixelIs set in the circuit (a) and the signal voltage of the circuit (b). Further, the calculation formulas of the response rate, the noise voltage, the noise equivalent temperature difference and the linearity characteristics generated in the step S2 are respectively expressed as follows: In the above-mentioned method, the step of, Representing picture elements at different temperaturesIs used for the response rate of the (a),、Pixels with temperature T in high-illumination scene and temperature T0 in low-illumination scene respectivelyA signal voltage; Representing a picture element in a dark field Is set to be a noise voltage of a (c),Representing a picture element in a dark fieldIs used for the response voltage of the (a),Representing the f-th frame pixel at the low illumination scene temperature T0Is set to the signal voltage of (1); Representing the noise equivalent temperature difference; A non-linear deviation representing the characteristic of the linearity, Representing picture elementsThe original signal voltage at temperature T in a high light scene,、Representing picture elementsResponse curve fitting coefficients of (a). Further, the step of constructing a blind pixel table based on blind pixel detection parameters specifically comprises the steps of fusing a spatial response criterion, a response rate criterion, a time domain noise criterion, a linearity criterion and a G parameter criterion, generating the blind pixel table through logical OR operation, wherein when the blind pixel table is used for carrying out blind pixel marking on an image to be a