JP3531256B2 - Automatic correction method for variation of infrared detector - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an infrared image pickup device which receives infrared rays from an object and creates an image of a surrounding scene of the object, and more particularly to a two-dimensional array type infrared detector which is an image pickup element. Variations in characteristics of each element (characteristics of electrical output with respect to optical input) (becomes noise of fixed pattern)
The present invention relates to an automatic correction method for variations in infrared detectors to improve the deterioration of the image quality of the target scene due to unnecessary signals generated by.

[0002]

2. Description of the Related Art When an infrared detector of a two-dimensional array scans a surrounding scene of an image pickup object and obtains a two-dimensional image signal by an electric output with respect to an optical input, it is unnecessary due to variations in input / output characteristics of each element. As shown in the configuration diagram of the conventional example of FIG. 4, the video signal obtained by digitizing the output signals of all the elements including the signal is used as the input signal Dip. First, in the frame integration section (1), the two-sided configuration ("f = 0 The integrated output Sf′p (“f = 1” plane) of the frame one before the current frame Sfp (“f = 0” plane) from the frame memory (4) of the RAM of the “plane and“ f = 1 ”plane) ) As the reference and t as the time constant of the frame integration, the integrated output Sfp of the current frame is calculated by using the input signal Dip and the integrated output Sf′p of the previous frame, and the calculation formula Sfp = Sf′p + ( Dip-S
f′p) / t, that is, the input signal data Dip of the outputs of all the elements including the defective element are subjected to frame integration processing for sequentially smoothing them temporally, and the processed output data Sf ′ p or Sfp is used as the calibration data Cp for correcting the unnecessary signal due to the variation of the input / output characteristics of each element in the input signal Dip, and the correction calculation unit (3) inputs the input signal Di
A correction method in which the effect of variations in the input / output characteristics of each element is corrected by adding to p and the output image data Dop is output is well known in the art.

[0003]

In this conventional correction method, an image of a surrounding scene including variations obtained by scanning an inhomogeneous surrounding scene by an infrared detector of a two-dimensional array is temporally smoothed. In the calibration data obtained,
In addition to the variations required for the calibration data, spatial low-frequency signals due to surrounding scenes that are unnecessary as calibration data remain. Therefore, when the unnecessary signal is corrected by the variation of each element using the calibration data, the distribution of the unnecessary signal in the calibration data is superimposed on the output video signal obtained after the correction, and the surrounding scene However, there is a problem in that a normal video signal cannot be obtained. An object of the present invention is to perform temporal smoothing processing on a digital video signal input Dip including an unnecessary signal due to variations of each element obtained by scanning a surrounding scene by an infrared detector of a two-dimensional array. In the variation correction method of the infrared detecting element that corrects the processed output by adding it to the input signal Dip as the calibration data, the influence of the unnecessary signal in the digital video signal corresponding to the infrared intensity distribution of the surrounding scene is The goal is to ensure that nothing remains in the calibration data.

[0004]

The basic configuration of the correction method of the present invention for achieving the above object is as follows. The infrared detector is configured to perform a correction calculation for the output having variations in each element. That is, for the digital signal input Dip of the two-dimensional frame image, which is the output signal with the variation of the input / output characteristics of each element of the infrared detector of the two-dimensional array, the integration process [Sfp = Sf ′ p + (Dip-Sf ′ p) / t] forever and outputs the processed output Sfp (1)
Then, the processed output (Sfp or Sf′p) is processed by a spatial filter that defines the surrounding area, and the output Sf′p of each element of the processed output and the average value Rf′p of the output of the peripheral elements are processed. Difference with (R
f'p-Sf'p = Cp) to generate the calibration data Cp, and the calibration data Cp of the output and the digital image signal input Dip are added to the characteristics of each element. Compensation calculation unit (3) that corrects the variation of
The correction calculation unit (3) is configured to output the corrected correct output image data Dop.

[0005]

In the present invention, referring to the operation explanatory view of FIG. 2, first, due to variations in the input / output characteristics of each element from the infrared detector of the two-dimensional array (not shown), unnecessary signals of positive or negative polarity are generated. For the scenes (A) to (D) in FIG. 2, which are digital video signal input Dip that are generated and exist at a fixed position regardless of the scanning of the surrounding scenes, the frame integration unit (1)
At [Sfp = Sf ′ p + (Dip-S
f′p) / t] is performed to obtain the output Sf′p of the temporal smoothing processing shown in (E) of FIG.

Next, the processing output Sf ′ obtained above is obtained.
Using p, the difference between the output Sf′p of each element that defines the surrounding area as a spatial filter and the average value Rf′p of the output of the peripheral elements (Rf′p − Sf ′ The calibration data Cp of the new method according to the present invention is obtained as p = Cp). Figure 2
In (G), the process of obtaining the average value Rf'p is performed by the spatial filter of gain = 1/8 for the surrounding 8 elements, and the calibration data Cp for the input signal Dip which is the output from each element is created.

Finally, in the correction calculation section (3), the digital image signal Dip output from the detection element is shown in FIG.
By adding the calibration data Cp obtained in ([G] Calibration data Cp of the new method in Fig. 2] to the output Dip from the (H) sensing element of, the variation in the characteristics of each element is corrected. , (J) Obtain the correction result Dop of the new method in FIG. Therefore,
In the present invention, based on the video signal Dip of the actual scene including the unnecessary signal due to the variation in the characteristics of each element obtained from the infrared detector of the two-dimensional array, the calibration data Cp is created and the variation is corrected, Highly accurate correction that matches the infrared intensity of the surrounding scene that is always imaged is performed. In addition, (F) of FIG. 2 is the calibration data Cp of the conventional method, and it is necessary to correct the positive / negative of the output of the individual element with respect to the average value 0 of the outputs of all the elements by the calibration data Cp of the opposite polarity. 2 (I) shows the correction result Dop of the conventional method obtained by adding the output Dip of (H) to the correct correction output (shown by the dotted line) and the distorted correction output (solid line). (Indicated by ─).

[0008]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 3 shows the configuration of an automatic correction unit for variations in an infrared detecting element according to an embodiment of the present invention. The digital input signal Dip is a digital signal of the output signal of the infrared detector of the two-dimensional array.It is a digital signal that also contains defect information of each element, and a digital image signal input synchronized with various sync signals. Is. Address generation circuit (30)
R inputs the various synchronization signals thereof and pre-stores the information of the defective element corresponding to the digital image signal input Dip.
A frame integrator (4) that integrates, for each frame, the output data Dp obtained by adding the defect information to the address for reading the data output of the OM (20) and the digital image signal input Dip.
The output data Sfp from (0) is stored as a data input, and the address for reading the data output for obtaining the input data Sf′p to the frame integration unit (40) in the calibration data clear circuit (60) is set. Occur. The ROM (20) that stores the information of the defective element in advance stores the information whether each element of the infrared detector is normal or defective, and the defect corresponding to the input of the address from the address generation circuit (30). The information (0 for normal and 1 for defect) is output to the defect information addition unit (10). The defect information adding unit (10) processes the defect information from the ROM (20) of the defect information so that the value of the normal pixel is 1 or more and the value of the defective pixel is 0, and outputs the output data Dp. . By doing so, the pixel having the value 0 is defective in the subsequent frame integrator (40), calibration data generator (70), and offset corrector (80). The frame integrator (40) receives the output data Dp to which the defect information is added and its own output Sfp up to the previous frame and stores the accumulated frame integration RA.
The output data Sf'p of M (50) is input and integrated. The formula for the integration is Sfp = Sf'p + (Dp-Sf'p) / t, p is the pixel number, t is the time constant of the integration, and f is the write frame number, 0 or 1, and each frame is Changes to 0 → 1 → 0 → 1. Then, the result of integration is sent back to the RAM (50) for frame integration as output data Spf. At that time, input data
The integration command signal, which is a signal synchronized with the Dip frame, is 1
When the above, the above integration is performed, and when 0, the output of RAM (50)
The Sf′p is returned as it is to the RAM (50) as Spf, and at the same time, it is output to the calibration data creating unit (70), and the calibration data creating unit (70) outputs the surrounding area defined as the spatial filter. The total value A of the output values of the pixels and the number of surrounding pixels other than 0, n = B, are calculated, the ratio A / B to each other is calculated, and the difference between A / B and Sf′p is calculated by a differencer. The calibration data Cp is created, and the calibration data Cp of the previous frame is updated or fixed. The RAM (50) for frame integration has a storage capacity of two frames ("f = 0" surface and "f = 1" surface) of the frame, and the address generation circuit (30)
The output data Sfp of the frame integrator (40) is stored by switching which one of them is the writing surface or the reading surface for each frame by the address from. The output Sfp of the current frame is written to the writing surface, and the output Sfp written to the previous frame is read as Sf′p from the reading surface. The clear circuit (60) is an AND gate,
Output of the frame integrator (40) by clearing (outputting) 1 of the output Sf′p from the frame integration RAM (50) by the calibration data clear signal 1 synchronized with the Dip frame.
Sfp will be output, and at the same time, the output Sfp will also be written on the writing surface of the frame integration RAM (50).
The output data Sf'p of the frame integration RAM (50) is output to the calibration data creation unit (70). Calibration data creation section
(70) inputs the data Sf'p stored in the frame integration RAM (50) up to the previous frame, calculates the calibration data Cp excluding the effect of defective pixels (calculation formula is shown in the figure), and outputs it. To do. First, the output values of the peripheral pixels of the pixel of interest are summed. At this time, the output value of the defective pixel has already been set to 0 by the defect information addition circuit unit (10), and the result of integration by the frame integration unit (4) is also 0. Adding all does not affect the defective pixels (thus the defective pixels are not totaled). At the same time, the number of surrounding pixels other than 0 is counted, which is the number of normal pixels obtained by removing the number of defective pixels from the number of surrounding pixels. By dividing the total output value of the surrounding pixels obtained in this way by the number of normal pixels of the surrounding pixels, the average value Rf'p excluding the defects of the surrounding pixels is calculated. Furthermore, from this average value Rf′p, the output Sf′p of the pixel of interest is
The calibration data Cp is calculated by subtracting the value of. The offset correction unit (80), which is the correction calculation unit (3) in the principle diagram of the present invention in FIG. 1, adds the correction data Cp to the uncorrected input data Dp to obtain the corrected data Dop. Is obtained and is output as data of an output image having no unnecessary signal due to variations in input / output characteristics of each element.

[0009]

As described above, according to the present invention, based on the video signal of the actual scene including the unnecessary signal due to the variation of the input / output characteristics of each element obtained from the infrared detector of the two-dimensional array, Since calibration data for unnecessary signals due to the variation is created and corrected by adding it to the video signal input, high-accuracy correction that always matches the infrared intensity of the surrounding scene will be performed, and the image of the infrared imaging device will be captured. An effect of improving image characteristics can be obtained.

[Brief description of drawings]

FIG. 1 is a principle diagram showing a basic configuration of an automatic correction method for variations of an infrared detection element according to the present invention.

FIG. 2 is an operation explanatory diagram of an automatic correction method for variations of an infrared detection element of the present invention.

FIG. 3 is a circuit configuration diagram of an automatic correction unit for variations of an infrared detection element according to an embodiment of the present invention.

FIG. 4 is a configuration diagram of a conventional method for automatically correcting variations in infrared detection elements.

[Explanation of symbols]

(1) is the frame integration part, (2) is the calibration data creation part, (3)
Is a correction calculation unit, (4) is a frame memory, (10) is a defect information addition circuit, (20) is a defect information ROM, (30) is an address generation circuit, (40) is a frame integration unit, and (50) is RAM for frame integration, (60) clear circuit, (70) calibration data generator,
(80) is an offset correction unit.

Front page continuation (72) Inventor Masaki Kamata 1333 Kamitadanaka, Nakahara-ku, Kawasaki-shi, Kanagawa Inside Fujitsu System Integration Laboratories, Ltd. (56) Reference JP-A-5-264357 (JP, A) JP-A-7-255014 (JP, A) Actual Kaihei 3-120178 (JP, U) (58) Fields investigated (Int.Cl. ⁷ , DB name) H04N 5/33-5/335 G01J 1/44 G01J 5/48

Claims (3)

(57) [Claims] 1. A method for automatically correcting deterioration of a picked-up image due to an unnecessary signal caused by a variation in input / output characteristics of each element of a two-dimensional array infrared detector which is an image pickup element of an infrared imager. Calibration data (Cp) that corrects the deterioration of the image due to unnecessary signals generated by the variation of the input / output characteristics of the element, the output (Sf'p) of each element in the area for each local area of the array and its periphery Difference from the average value (Rf′p) of the element output (Cp = Rf′p−
Sf′p) is created as an automatic correction method for variations of infrared detecting elements. 2. When operating the infrared imaging device,
The surrounding scene is scanned and imaged, and each sensing element at that time
Output signal (Dip)AgainstProcessing to smooth temporallyLine
IData after that processingFor each local area within that area
Output of each element of (Science fiction ′ p) And the average value of the output of the peripheral element
(Rf ′ p) Difference from (Cp = Rf ′ p − Science fiction ′ p) And calculate the calibration data
Data (Cp) To createThe red according to claim 1, characterized in that
A method for automatically correcting variations in external line detection elements. 3. The average value (Rf ′ p) of the outputs of the peripheral elements is
When calculating, the defective element is not included in the relevant peripheral elements.
If it is, calculate the average value excluding the defective element.
The method for automatically correcting the variation of the infrared detection element according to claim 1, wherein