JPS6267513A - A method for compensating for image deterioration in visual devices using an image fiberscope - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Summary] In a visual device that photoelectrically converts an image transmitted through an optical system including an image fiberscope using an array sensor, the output of the sensor is based on the characteristics of the optical system. The purpose is to improve the quality of images by repeatedly performing arithmetic processing.

[Industrial application field]

Images transmitted by an optical system such as the one shown in Figure 2, including an image fiberscope, have the disadvantage that the resolution is limited by the diameter of the fiberscope, so the image quality is inferior to that obtained by a camera or the like. there were.

Since the present invention improves the quality of images by processing electrical signal output from images obtained through such optical systems, the present invention improves image quality by processing electrical signal output from images obtained through such optical systems. It can be applied to equipment configured to obtain

[Conventional technology]

In the past, the image was usually improved by improving the image fiber optical system.

[Problem that the invention seeks to solve]

The resolution of images in visual devices using image fiberscopes is not only limited by the diameter of the fiberscope used, but also has the disadvantage of lacking image smoothness.

The present invention attempts to compensate for the deterioration in image quality by repeatedly processing electrical signals obtained by photoelectrically converting images obtained through an optical system including a fiberscope using an electronic computer. .

[Means for solving problems]

In order to correct the image signal obtained through the image fiberscope, the point light source (0) is imaged through the image fiberscope (F), and the photoelectric conversion sensor array is Point spread function h(
x, y).

Here, to simplify processing, we assume that the optical system including the image fiberscope satisfies the condition of isoplanatism without comatic aberration, h (x t )')
=h (x x'·yy'), and the point spread function can be regarded as constant regardless of the position on the screen.

Therefore, the output image y on the two-dimensional array sensor (S) is y (x, y)=ΣΣh(x x/, y y')・
It will be expressed as x (x', y'). Here x (x' s y
) is the true value of the image at the (x'ty') point.

Based on this premise, the present invention compensates for image deterioration by applying a learning identification method.
, n are the coordinates of pixels on the X-axis and y-axis, respectively, and j is the number of repetitions of arithmetic processing, then the j-th calculation result zJ is corrected with respect to the previous calculation result Z shall be.

Here e j−+ = V j−14s+ +11 ”j−
+ (an II)Wj-+ +++i+n) -()
'+ ++m- to +fi-k), Zj-1++a- to +
r+-Ll), lJ is the j-th modified image, and h is the point spread function mentioned above, both of which are known, so if we set the coefficient μ to 2, then lJ can be calculated. By repeating the process, it converges and the desired image can be obtained.

Geometrically explaining the reason for convergence by setting μ within the above range in a three-dimensional case, we can say that h, (-h)
, Zj, and M are expressed as shown in FIG. 4 when μ=1. The plane Sj is a hyperplane where (h-Zj)=Wj holds, and the plane S,-1 is a hyperplane where (h-Z,-1)=Wj-1 holds.

In the present invention, Z j −+−
2, it attempts to approach the true value successively, and Δ
2 is a correction value for the previous value mentioned above.

If μ=1, ΔZ//hj holds true, so Zj is the position of the foot of the perpendicular line drawn from Zj-+ to plane SJ, and if μ≦0, 1lZj-Me11≧11Z J −
+ x II, if μ≧2 then 1lZJ−xll>I
If IZ, -+ xll holds true, and in the end it is 0, then 0, then 1IZ4-x If < If ZJ-+ -x I
I Tona Lucaratial.

〔Example〕

FIG. 1 shows an embodiment of the arithmetic processing device of the present invention.
The initial value or previous value lJ, , 1 read from the memory (1) is multiplied by the point spread function h and the multiplier (2), and then the array sensor (S)
is subtracted from the image output y from memory (1), amplified and added to the previous value Zj-+ from memory (1) to form a new value lJ, updating the previous value 14-1 in memory (1).

The number of times this calculation process is repeated can be determined by providing a counter to count the number of calculations and stopping the calculation when it reaches a preset number of times, or by monitoring the image obtained from the image signal to obtain an appropriate image. When the value is obtained, the calculation may be stopped according to an instruction from the operator.

〔Effect of the invention〕

According to the present invention, an image obtained through an image fiberscope can be smoothed, so that the image becomes significantly easier to see.

[Brief explanation of drawings]

Fig. 1 is an embodiment of the arithmetic processing device of the present invention, Fig. 2 is a principle diagram of a visual device using an image fiberscope, Fig. 3 is a diagram explaining the point spread function of this visual device, and Fig. 4 is a diagram illustrating the point spread function of this visual device. FIG. 3 is a diagram for explaining the range of coefficient μ. *XP Kasai 6J resistance water moat

Claims (1)

[Claims] In a visual device that photoelectrically converts an image transmitted through an optical system including an image fiberscope using an array sensor, the output of this visual device is Image (y
) and the point spread function (
h) and the scalar product of the appropriately set initial value or the value (Z_j_−_1) obtained from the previous calculation process,
To prevent image deterioration in a visual device using an image fiberscope, which is characterized by repeating arithmetic processing in which a value obtained by multiplying this difference value by a constant is added to the above-mentioned initial value or a value (Z_j_−_1) obtained from the previous arithmetic processing. Compensation method.