JP2001133223A - Method of correcting distortion aberration - Google Patents

Abstract

PROBLEM TO BE SOLVED: To highly accurately measure an image. SOLUTION: Image heights h1-h5 based on actual measurement and the image heights h1'-h5' on a picked-up image corresponding to them are respectively obtained and a distortion aberration rate is obtained based on the values by the equation, the distortion aberration rate (%)=100×(h'-h)/h. When such a distortion aberration rate is successively obtained along X' and Y' axes, the correction curve of the distortion aberration rate is obtained. When the correction curve is obtained beforehand, a measured value at the time of picking up the image of a measurement object by a digital camera can be applied to the correction curve and corrected.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a method for correcting distortion, and more particularly to a method for improving measurement accuracy in image measurement.

[0002]

2. Description of the Related Art Image measurement using a digital camera such as a CCD is used for inspection and processing management of industrial products, and is mainly used in fields where not so high measurement accuracy is required.

[0003] By the way, since this kind of image measurement can be measured in a non-contact manner, for example, measurement of displacement over time of a structure such as a bridge or the like can be performed with high precision in a part that is out of reach or a dangerous part. It is expected that it will be used for measurement.

[0004] However, there are the following technical problems in applying the conventional image measurement using a digital camera to a field requiring high measurement accuracy.

[0005]

That is, in the conventional image measurement, measurement points are extracted from captured image data obtained by imaging, the number of pixels between the measurement points is counted, and the pixel count is added to the obtained count value. The length between measurement points is obtained by multiplying the size.

However, although digital cameras such as CCDs have been improved in accuracy by increasing the number of pixels, etc., they have a lens system, and the captured image is affected by the aberration of the lens system. The receiving part is bent into a barrel shape or a pin winding shape, and if this is counted as it is, since it includes an error due to aberration, there is a problem that a highly accurate measurement value cannot be obtained.

SUMMARY OF THE INVENTION The present invention has been made in view of such conventional problems, and an object of the present invention is to provide a method of correcting distortion which enables high-precision image measurement. is there.

[0008]

In order to achieve the above-mentioned object, the present invention provides a digital camera which images a calibration plate on which a plurality of target marks whose coordinate values are known are displayed, and displays the image on a screen. Based on the measured values between the target marks and the known coordinate values, create a distortion aberration correction curve corresponding to the shooting distance of the calibration plate, based on the obtained distortion aberration correction curve,
Corrected the measured value. According to the distortion correction method configured as described above, the distortion aberration correction curve corresponding to the shooting distance of the calibration plate is calculated based on the measured values between the target marks on the captured screen and the known coordinate values. Based on the created and obtained distortion aberration rate correction curve,
Since the measured value is corrected, highly accurate measurement can be performed. Further, the present invention captures a calibration plate on which a plurality of target marks having known coordinate values are displayed by a digital camera, and based on the measured values between the target marks on the captured screen and the known coordinate values. Then, the parameters of the multidimensional polynomial correction approximation equation are determined, and the corrected correction value for the measured value is obtained from the obtained multidimensional polynomial correction approximation equation. With such a configuration, distortion correction can be performed without depending on the imaging conditions of the digital camera.

[0009]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Preferred embodiments of the present invention will be described below in detail with reference to the accompanying drawings. 1 to 4 show one embodiment of a method for correcting distortion according to the present invention.

FIG. 1 shows a calibration plate C used in the correction method of this embodiment. The calibration plate C shown in the figure has a total of five target marks TA to TE accurately drawn at the positions of the dimensions shown. The target marks TA to TE are formed by drawing a perfect circle of a predetermined diameter in black on a white paperboard, and are in a state where an image can be taken by the digital camera 14.

When obtaining the distortion, a plurality of calibration plates C are arranged vertically and horizontally on the vertical wall at predetermined intervals, as shown in FIG. This is imaged by the digital camera 14 from a point at a predetermined shooting distance Z.

Each calibration plate C arranged on the wall
Are the two-dimensional coordinates of the X, Y system having the origin at the target mark TA _{i, j} of the central calibration plate C _{i, j} arranged in m columns and n rows as shown in FIG. The value is measured with a laser transit having an error of about ± 1 mm.

The laser transit is measured by a level, and placed on a tripod 16 so as to be parallel to the ground. The digital camera 14 is mounted on the same tripod 16 and adjusted so that the optical axis is directed in the direction of the target mark TA _{i, j} of the central calibration plate C _{i, j} , so that the calibration plates C having the same arrangement are arranged. Take an image.

FIG. 4 shows the main part of the image screen when the target C is imaged by the digital camera 14. In the image captured by the camera 14 _, the coordinate values of each target mark TA ′ _{i, j and the} like are obtained from the X ′, Y ′ two-dimensional coordinates having the origin at the center of the pixel p _ij .

In this case, the size of the target marks TA to TE on the imaging screen is, for example, 6 million pixels of CC.
When shooting at a shooting distance of about 15 m using a D camera, the distance between one pixel is about 4 m on the imaging screen from the angle of view of the lens.
m, so, for example, the target mark TA ~
When the size of TE is a circle of 50 mm, about 10 to 13
Occupies the size of a pixel.

Therefore, the center positions of the target marks TA to TE are set on the center of gravity of the circle on the imaging screen. In this case, since the target marks TA to TE are drawn in black, the position of the center of gravity is determined after clarifying the figure by performing threshold processing.

In this case, even if the digital camera 14 is arranged so that the optical axis is oriented in the direction of the target mark TA _{i, j} of the central calibration plate C _{i, j} , the center of the optical axis and the pixel p _ij are not aligned. A shift T as shown in FIG. 4 occurs between the center and the center, and the shift T becomes an error between coordinate systems.

The magnitude of the deviation T is determined by obtaining the coordinates of the target mark TA ′ _{i, j} of the calibration plate C ′ _{i, j} with the center of the pixel p _ij as the origin, and the magnitude of the deviation T is determined. Then, error correction between coordinate systems is performed to correct each coordinate value of the target mark TA ′ _{i, j and the} like.

When the error correction between the coordinate systems is completed, the image heights h1 to h5 based on the actual measurement shown in FIG. 3 and the corresponding image heights h1 'to h5' on the imaging screen shown in FIG. Each of them is obtained, and based on these values, the distortion aberration rate is obtained by the following equation.

Distortion Aberration Rate (%) = 100 × (h′−h) / h When such distortion aberration rates are sequentially obtained along the X ′ and Y ′ axes or the radiation direction, FIG. Such a correction curve of the distortion aberration rate is obtained. FIG. 5 shows three images obtained by imaging the calibration plate C with a CCD camera having 6 million pixels each at a shooting distance of 15 m, 10 m, and 5 m.
Different types of correction curves are shown.

In FIG. 5, the horizontal axis represents the image height after the distortion, and it can be seen that there is no large deviation compared to the theoretical value except for the portion above the image height of 11 mm at a shooting distance of 5 m. Was.

If such a correction curve is obtained in advance, it is possible to correct a measurement value obtained by imaging the object to be measured by the digital camera 14 by applying the correction curve to the correction curve. When used, the accuracy of image measurement can be improved. Incidentally, the correction curve of the distortion aberration rate obtained as described above differs depending on the imaging distance of the digital camera 14, and changes when the direction of the angle of view of the digital camera 14 is directed to a different position. When employed for distortion correction, there is a problem in reproducibility.

Therefore, a correction curve can be approximated by a multidimensional polynomial as a method of correcting distortion that is independent of the imaging distance without being affected by these conditions.

In this embodiment, the following five-dimensional polynomial correction formula is set as a function of converting the image height h 'after distortion into the image height h before distortion. h = ah ' ⁵ + bh' ⁵ + ch ' ⁴ + dh' ³ + eh 'In order to obtain each parameter of this correction formula, five or more measurement points are set on the imaging screen shown in FIG. The image heights h1 'to h5' at each measurement point and the corresponding actually measured image heights h1 to h5 shown in FIG. 3 are respectively obtained, and five or more equations obtained by substituting these values are obtained. Are used to determine the parameters (a to e) at which the most approximate values are obtained.

FIG. 6 shows an actual calibration plate C
At a shooting distance of 15 m with a 6 million pixel CCD camera, parameters (a to e) are determined by the above method, and distortion is corrected. The error between them is shown.

As is clear from the results shown in FIG.
If the distortion is not corrected, the error from the center is about 6 m, and an error of about 100 mm occurs. However, if the aberration correction is performed, the error falls within ± 10 mm.

According to the distortion aberration correction method as described above, the parameters of the multidimensional polynomial correction approximation equation are determined in advance based on the measured values between the target marks on the imaged screen and the known interval values. Since the corrected coordinate value subjected to the distortion correction is obtained from the multidimensional polynomial correction approximation formula, the absolute distortion correction independent of the imaging conditions of the digital camera 14 can be performed.

[0028]

As described above in detail, according to the distortion correcting method of the present invention, highly accurate measurement can be performed.

[Brief description of the drawings]

FIG. 1 is a plan view of a calibration plate used in a distortion aberration correcting method according to the present invention.

FIG. 2 is an explanatory diagram when photographing the calibration plate shown in FIG. 1 with a digital camera.

FIG. 3 is an explanatory diagram of a main part of actually measured coordinate values of the calibration plate shown in FIG. 2;

FIG. 4 is an explanatory diagram of a main part of coordinate values on an image of the calibration plate shown in FIG. 2;

FIG. 5 is a graph of a correction curve used for correcting distortion according to the present invention.

FIG. 6 is a graph showing errors when distortion correction of the present invention is performed and when distortion correction is not performed.

[Description of Signs] TA to TE Target Mark C Calibration Plate 14 Digital Camera

 ──────────────────────────────────────────────────続 き Continuing from the front page (72) Inventor Ichiichi Ouchi 4-640 Shimoseito, Kiyose-shi, Tokyo Inside the Obayashi Corporation Technical Research Institute (72) Inventor Shuji Hashimoto 1-13-19 Yawata, Ichikawa-shi, Chiba F-term (Reference) 2F065 AA04 AA53 BB28 CC14 EE08 FF04 FF61 JJ03 JJ26 QQ17 RR10 2F112 AC06 BA20 CA08 CA12 DA40 FA23 FA35 5B057 AA01 BA02 BA11 CA12 CA16 DA07 DA20 DB02 DC03 5C022 AA01 AB51 AC42 AC76

Claims (2)

[Claims] An image of a calibration plate on which a plurality of target marks having known coordinate values are displayed is captured by a digital camera, based on a measured value between the target marks on the captured screen and the known coordinate values. A distortion correction curve corresponding to the shooting distance of the calibration plate, and correcting the measured value based on the obtained distortion correction curve. 2. A digital camera captures an image of a calibration plate on which a plurality of target marks whose coordinate values are known are displayed, based on a measured value between the target marks on the captured image and the known coordinate values. A method for correcting distortion, comprising: determining parameters of a multidimensional polynomial correction approximation equation; and obtaining a correction correction value for a measured value from the obtained multidimensional polynomial correction approximation equation.