JPH0362213B2 - - Google Patents

Description

[Detailed description of the invention] Table of contents [Summary] ...2 pages [Industrial application fields] ...3 pages [Prior art] ...3 pages [Problems to be solved by the invention] ...6 pages [ Means for Solving the Problem] ...Page 7 [Operation] ...Page 8 [Example] ...Page 9 [Effects of the Invention] ...Page 24 [Summary] A lens projection characteristic measuring device, which has a circular shape. By making the pattern light incident on the lens by reflection from a mirror, it was possible to simplify the device, improve accuracy, and make it possible to measure a wide range of lenses.

[Industrial application field]

The present invention relates to a lens projection characteristic measuring device for measuring the projection characteristic of a lens, and more particularly to a lens projection characteristic measuring device using annular pattern light.

[Conventional technology]

2. Description of the Related Art In recent years, technology for environmental recognition using a camera device consisting of a lens and an image sensor as an image input means has become important. In such technical fields,
Accurate environmental recognition cannot be performed without knowing the projective characteristics of the lens.

For example, as shown in FIG. 12, it is known where on the imaging surface 23 the light 22 incident on the lens 21 from a certain direction is imaged, and based on this measurement, the location where the image is formed is determined. This is because the direction of incidence of light is recognized.
In the example shown here, the light incident from the extra latitude φ and the azimuth θ is focused on a point P (X, Y) on the imaging plane 23. These (φ, θ) and (X, Y) are related by the theoretical projection characteristics of the lens, but in reality, it is difficult to measure the projection characteristics of the lens due to the presence of errors. It becomes necessary.

Conventionally, lens aberrations have been widely measured in order to determine lens characteristics, and devices that automatically measure lens aberrations have also been developed. However, no effective device has yet existed as a means for measuring the projection characteristics of a lens.

Although there is no device that can measure the projection characteristics of a lens, the lens is irradiated with light from each direction and this image formation is electrically detected by an image sensor on the imaging surface, and the direction of the light incidence and the image formation are measured. Various devices can be envisaged as means for measuring position and relationship.

However, in the various examples of such assumed devices, the supporting devices and the like become large in order to move the heavy television camera, and it is difficult to increase the precision of the angle control of the incident light.

Conversely, it is possible to move the collimator on the light irradiation side, but the collimator is also heavy like a television camera and has the same problems.

In order to solve these problems, the present inventors have already proposed a ``device for measuring projection characteristics of lenses'' (Japanese Patent Application No. 125249/1982). In order to solve the above problems and obtain a compact and highly accurate lens projection characteristic measuring device, this device includes an irradiation device that directs light from a predetermined direction toward the lens of the imaging system; In a lens projection characteristic measuring device that includes a measuring device that measures where light is focused on the imaging surface, the irradiation device is tilted with respect to the optical axis of the lens and measures a point on an upwardly extending line of the optical axis. It is composed of a projector that emits light toward the lens, a mirror that is movable along the direction of the light and whose mounting angle is variable and reflects the light toward the lens, and a rotating table that rotates the lens.

In this measuring device, the light from the projector is reflected by a mirror and enters the lens, but by changing the position and angle of the mirror, the co-latitude of the incident light relative to the lens can be changed, and the lens can also be rotated. By rotating it on a stand, light can enter the lens from any azimuth angle.

[Problem that the invention seeks to solve]

However, in the case of the existing proposal, when measuring the geometric relationship between the angle of incident light on the lens and the angle of transmitted light, and finding the center of the light transmitted through the lens to be measured, cross slit light is used as the incident light. ing.

That is, as shown in FIG. 13, the irradiation light 35 that enters the lens to be measured 33 from the autocollimator 31 serving as a projector is a cross beam that passes through the cross slit section. Transmitted light 3 of the lens to be measured 33
7 is projected on the screen 39. On the screen 39, there is a cross image 41 as shown in FIG.
obtained as.

The cross-shaped image 41 thus obtained is convenient for finding the intersection O ₁ with the human eye. However, if it is automated to achieve high accuracy, the 14th
As shown in figure a, the center of gravity of the cross image 41 must be set at the intersection _O1 , which makes it vulnerable to noise during detection.
It was necessary to cut out only 1 for detection, and it was extremely difficult to automatically determine the center of transmitted light based on the detection data.

Also, as shown in FIG. 14b, a straight line l _x
Even if calculations were performed in which the intersection point O ₂ was applied as the center of light transmitted through the lens, the amount of calculation would be enormous and automation would be difficult.

The present invention was created in view of these points, and an object of the present invention is to provide a lens projection characteristic measuring device that can automatically determine the center of light transmitted through a lens to be measured with high accuracy.

[Means for solving problems]

In order to achieve the above object, the present invention includes a light irradiation means 115 capable of injecting light from a predetermined direction toward the lens of the imaging system, and a measurement means 121 that measures where on the imaging surface this light is focused. In the lens projection characteristic measuring device, the light irradiation means 1
15 has a plurality of concentric pattern light irradiation functions 113 which are circular slit parallel lights to the lens to be measured 111 of the imaging system, and the measuring means 121
determines the optimal circular image whose center should be determined from among the plurality of received circular images when determining the center of the circular image on the imaging surface based on the received light data of the plurality of concentric circular pattern lights. It is configured to have a function.

[Effect]

In the present invention using the above means, first, the irradiation means makes a plurality of concentric parallel beams of light enter the lens to be measured. Then, multiple circular images appear on the imaging surface. Then, the function of the measuring means 121 automatically selects the slit light most suitable for determining the center from among the plurality of circular images, thereby determining the center position from the optimal circular image. This provides a highly accurate lens projection characteristic measuring device.

〔Example〕

Embodiments of the lens projection characteristic measuring device according to the present invention will be described below.

In this embodiment, a television camera 212 to which a lens to be measured 211 is attached is attached to a rotary table 213 so that it can rotate around the optical axis A of this lens 211. An autocollimator 214 as a projector that irradiates the reference pattern as parallel light L ₁ is installed obliquely within a plane that includes the optical axis A of the lens 211, and the irradiated light is transmitted to the lens 211.
It is arranged to face a point on the upward extension of the optical axis A of 11. A moving shaft 215 is fixed approximately parallel to this irradiation light _L1 , and a mounting member 2 that moves along this moving shaft 215 is fixed to the moving shaft 215.
A mirror 217 is provided with a variable angle via 16.

FIG. 3 shows an embodiment of the present invention in more detail, and FIG.
The figure shows its control system.

The angle of the mirror 217 can be changed by a mirror rotation motor 220 and a speed reduction device with an accuracy smaller than the resolution of the image sensor, and the mirror 217 can be moved around a movement shaft 215 by a mirror movement motor 221. Further, the television camera 212 can be rotated by a camera rotation motor 222 in a rotating table. These movements such as rotation are controlled by a control section 227 as shown in FIG. That is, each motor 2
The movements of the motors 20, 221, and 222 are driven by the control unit 227 via a motor driver 228, while the movements of each motor 220, 221, and 222 are input to the control unit 227 by rotary encoders 223, 224, and 225 and a counter 226. The control unit 227 controls each motor 220, 221, 222.
Each motor is controlled based on the motion state of the motor and the image signal from the measuring section 230 connected to the television camera 212. In addition, the reference numeral 218 in the figure is the base, 21
A stand 9 is erected on the base and supports each device.

The measurement unit 230 includes a memory unit 233 that receives image signals 231 from each television camera 212 and stores the data after analog-to-digital (A-D) conversion, and an image information signal (digital) from the memory unit 233. and an arithmetic unit 239 which receives 235 and provides an input instruction signal 237.

This calculation section 239 provides a motor instruction signal 241 to the control section 227 and also introduces an encoder information signal 243.

In order to measure the projection characteristics of a lens with the lens projection characteristics measuring device according to this embodiment, light is made to enter a desired measurement position. That is, the extra latitude is determined by changing the position and angle of the mirror 217, and the azimuth is determined by rotating the rotary table 213. In this state,
The measurement unit 230 may measure where the image is focused on the imaging surface of the television camera 212.
Note that this measurement can also be performed automatically by storing the measurement position in the control unit 227 in advance.

Therefore, according to this device, the driving part is the mirror 217.
Since it is lightweight, the device is small and the mirror 21
In addition to improving the measurement accuracy of the lens projection characteristics by simply improving the accuracy of the angle change of 7 and the rotational operation of the television camera 212, the operation of each motor is automatically controlled. The projection characteristics of a lens can be measured extremely easily and accurately. Furthermore, by changing the position angle of the mirror, the extra latitude is 0 degrees (217 ₁ shown by the dashed line in Figure 2).
It can measure from 90 degrees to more than 90 degrees of extra latitude, and can also easily measure special lenses such as fisheye lenses.

By the way, the irradiation light _L1 according to the embodiment of the present invention is a circular slit light. That is, as shown in FIG. 5, the circular slit light from the slit light source 251 is converted into parallel light _L1 by the autocollimator 214. The light L _t transmitted through the lens to be measured 211 is also a circular pattern light, and the television camera 212
The image is taken at .

Image data captured by the television camera 212 is stored in the memory unit 233 as circular pattern data.

6a to 6d show each stage of processing performed by the measuring section 230. FIG. Moreover, FIG. 7 shows the order.

The process of determining the center of lens transmitted light will be described below.

First, the input instruction signal 237 is supplied from the calculation section 239 in the measurement section 230 to the memory section 233 .
At this time, the lens to be measured 211 receives the circular pattern light _L1 irradiated from the autocollimator 214.
A circular pattern 251 as shown in FIG.
The digital data is stored in the memory section 233 (step 311). In this case, data for one image is stored in the memory section 233, and the following processing is performed based on this stored data.

As shown in FIG. 6a, by scanning along the horizontal axis,
The abscissas of two points A and B that intersect with the circular pattern 261 are determined. From these two points A and B to the center point C
Calculate (=A+B/2). In this way, one screen is sequentially scanned in the horizontal axis direction (see Figure 6b), and each time the center point C ₁ , C ₂ , ..., _Co
(Step 312).

Next, all the center points C ₁ found in step 312
The vertical center line _l1 is determined by approximation by the least squares method based on ~C _o (step 313). By the way, some of the stored data contains noise-based data C _n1
may exist, but these effects are
It is removed in the process of calculating l ₁ .

In parallel with this, scanning is performed along the vertical axis to find two points that intersect with the circular pattern, and the center point is calculated in the same way (see Figure 6c).
Do this for the entire image (step
314).

Next, an approximated horizontal line _l2 is determined by the least squares method based on all the center point data determined by the vertical axis scanning (step 315).

Therefore, as shown in FIG. 6d, the center vertical line l ₁ and the horizontal line for the circular pattern 261
This means that l ₂ has been found.

The intersection O of these two lines _l1 and _l2 is determined from the approximate data (step 316). This intersection O is the center point of the transmitted light of the lens 211 to be measured.

In this way, a circular pattern of slit light
By using L ₁ as the irradiation light, it becomes possible to automatically determine the center point of the transmitted light of the lens 211 to be measured with high precision.

Incidentally, there are a wide variety of lenses 211 to be measured, and their focal lengths differ greatly.
Therefore, for irradiation light L ₁ with the same circular pattern,
Depending on the size of the field of view of the lens to be measured 211,
The center may not be found accurately. That is, in order to accurately determine the center of a circle, for example, when an input image is quantized into 512×512 pixels as imaging data, the diameter of the circle needs to be about 10 pixels or less. If the circle is larger than this, the shape may be distorted due to the projection characteristics of the lens, and the center may not be accurately determined.

Furthermore, if the lens to be measured 211 is a fisheye lens or a telephoto lens, the field of view may differ by a factor of 30 or more. For example, a circle with a diameter of 10 pixels using a fisheye lens with a field of view of 180 degrees becomes 360 pixels in diameter with a telephoto lens with a field of view of 5 degrees, so part of the circle may protrude from the screen, making it difficult to determine the position of the center of the light transmitted through the lens. become unable.

Conversely, a circle with a diameter of 10 pixels when using a telephoto lens with a field of view of 5 degrees becomes less than 1 pixel in diameter when using a standard, wide-angle, or fisheye lens with a field of view of 50 degrees. It is difficult to stably extract an image with a diameter of one pixel or less unless the light source is very bright, and the autocollimator 214 used as the light source of the lens projection characteristic measuring device cannot have high brightness. Furthermore, when using a television camera 212 that uses a solid-state image sensor as a light-receiving element, even if an image with a diameter of one pixel or less is sufficiently bright, it will fall between adjacent light-receiving windows, and there will be no part of the image in the image. Sometimes it doesn't appear. Therefore, in order to stably extract a circular image, its diameter must be at least 10
A pixel size is necessary.

Therefore, in the embodiment of the present invention, a plurality of concentric slit parallel beams are used as the light incident on the lens to be measured 211, and the optimum circular pattern of slit beams is automatically selected from among them.

Autocollimator 214 in the embodiment of the present invention
As shown in FIG. 8, the optical system is arranged so that the slit surface 333 of the slit 331 of the slit light source 251 corresponds to the focal length of the autocollimation lens 335. Since the viewing angle of the slit pattern is determined by this arrangement, the diameter of the slit 331 is determined.

An example of a slit pattern determined from this viewpoint is shown in FIG. Here, six circular slits are prepared, and the diameter of the smallest first circle ₃₃₁₁ is
The viewing angle is selected to be 0.1 degree, which is 1/50 of a lens with a field of view of 5 degrees. 1st
The diameter of the second circle 331 ₂ , which is larger than the circle 331 ₁ , is chosen to be 0.2 degrees so that the viewing angle is doubled for the same lens. Similarly, the third
The diameters of the circles 331 ₃ to 6th circles 331 ₆ are selected to be 0.4 degrees, 0.8 degrees, 1.6 degrees, and 3.2 degrees, respectively, so that the viewing angle is doubled.

This allows the field of view to be seen from a telephoto lens with a field of view of 5 degrees.
It has a slit pattern that can be commonly used for up to 180 degree fisheye lenses. For example, the outermost 6th circle 3
With a diameter of ₃₁₆ , the viewing angle is approximately 1/50 of a lens with a field of view of 180 degrees, and the diameter is approximately 10 pixels for an image size of 512 x 512 pixels. In this way, by doubling the diameters of the concentric circles of the slit 331, the circular image diameter in the input image can be adjusted to 5 to 10 pixels for any lens with a field of view from 5 degrees to 180 degrees. It is possible to ensure that there is always one circle of degree.

FIGS. 10a to 10d show circular data as imaged data when the control system shown in FIG. 4 performs processing for determining an optimal circular slit. FIG. 11 is a flowchart showing the order of the processing.

First, 6 slits based on the 6 slits shown in Figure 9.
Irradiation light _L1 having a circular pattern as a bundle of two slit lights is incident on the lens 211 to be measured, and the image data of the circular pattern light is stored in the memory section 233 as one image. Subsequently, the calculation unit 239 reads the image information signal 235 from the memory unit 233 and traces the circle for each scanning line in order to find the optimal circle.

(i) When the optimal circle is found Range 3 as one screen as shown in Figure 10a
40, read the circular data in the first scan line 1 (step 411),
Determine whether or not there are cuts A and B (step
412), if the judgment is positive, the distance between both cuts A and B is n
(=approximately 15) It is determined whether or not the number of pixels is less than (approximately 15) (step 413).

In step 413, an affirmative determination is made even if there is only one cut point, that is, the circular data and the scanning line intersect.
In the case of such a positive judgment, the left and right sides of both outermost cuts are defined as A and B (step
414). For example, as shown in FIG. 10b, in the first scanning line, there may be two or more sets of both cuts within n pixels, so the outermost two cuts are used.

Following step 414, the cuts of the locus are similarly determined for the next second scan line (step 415), and for both cuts A and B in that case, the left cut A is the first
It is determined whether the right cut B has moved to the left and to the right compared to the case of the scanning line (step 416). If affirmative, the cuts are sequentially detected using each scanning line up to n lines ahead (step 417), it is determined whether or not both cuts A and B become one point (step 418), and furthermore, both cuts A and B are detected as one point. and B is the screen range 340
It is determined whether or not it protrudes from the circle (step 419), and it is determined whether the interval between the circular data where the two cuts A and B that do not protrude are the most distant is equal to or less than n (step 420).

As a result, the distance between both cuts in the fourth scanning line is n
pixel or more, and both cuts on the 9th scanning line are less than or equal to n pixels, then the cut on the 10th scanning line becomes 1 point, so this becomes the optimal circular pattern P ₀ , and according to that circular pattern P ₀ The data of the circular slit light is used to measure the center of the transmitted light (step 421).

Based on the imaging data of the circular slit light employed in step 421, two orthogonal least squares approximation straight lines l ₁ and l ₂ (see Fig. 6 d) are created by the processing procedure already described with reference to Figs. ) to find the center point O of the circle 261 (P ₀ ), and find the center point of the light transmitted through the lens 211 to be measured (step 422).

The above is the processing order when it is appropriate to set a plurality of circular patterns in a bundle in advance.

(ii) When the interval between the cuts in the first scanning line is too large. In this case, a negative determination is made in step 413, so the cuts are detected using a scanning line several lines ahead from that line (step 423). , it is determined whether there are two cuts (step 424), the two detected cuts are defined as A and B (step 425), and then the process moves to step 417 to execute the same process as described above.

Here, for example, the fourth scan line is detected and executed in step 423.

If it is determined in step 424 that there is no cut yet, it is determined whether all scanning lines have been completed (step 426), and the process proceeds to the next scanning line (step 427), where a new circle cut is detected. It is determined whether or not it has been performed (step 428). If the determination is negative, the loop operation returns to step 426, and at step 428, where the cut of the circular locus is detected, an affirmative determination is made, the loop is exited, and the process returns to step 423, where the same execution process as described above is performed.

(iii) When the distance between both cuts does not become large In this case, a negative determination is made in step 416,
It is determined whether the left cut A moves to the right and the right cut B moves to the left (step 429). In case of negative judgment,
Furthermore, it is determined whether there are two other cuts within n pixels (step 430), and if there are two cuts in a set, the left side is defined as cut A and the right side is defined as cut B (step 431), and the process proceeds to step 416. Return to and execute the process described above.

If the judgment in step 430 is negative, proceed to step 426.
to move to.

(iv) In case of error For example, as shown in Figure 10c, both cuts A and B
If both move to the right or left, the combination is incorrect, so a suitable combination is searched after the determination in step 429. However, if an affirmative determination is made in step 429, for example, FIG.
In the case shown in step 432,
The center of the slit pattern is not on the screen, and an error message is output.

Also, each scan line has 2 cuts based on circular data.
It may not be detected at all. For example, if no cut is detected in the first scan (negative determination in step 412),
Furthermore, if no cut is found in the loop from steps 426 to 428, it is an error and the steps are still repeated.
Shifts to 432 and displays an error.

If an error occurs, correct the angle of the mirror or camera as necessary and start over.

With this configuration, the center of transmitted light of the lens to be measured 211 can be stably determined regardless of the size of the field of view of the lens to be measured 211, and the accuracy of measuring the projection characteristics of the lens can be improved. Furthermore, there is no need to replace the slit light source each time depending on the lens 211 to be measured.

〔Effect of the invention〕

As explained above, according to the present invention, since the annular pattern light is used as the incident light to the lens, the center of the lens transmitted light can be determined automatically and with extremely high precision. In particular, if a plurality of annular pattern lights are used and the annular pattern light suitable for the lens to be measured is used for measuring the center of the transmitted light, the measurement can be performed under optimal measurement conditions.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the basic structure of a lens projection characteristic measuring device according to the present invention, FIGS. 2 and 3 are block diagrams showing an embodiment of the present invention, and FIG. 4 is a block diagram showing a control system in an embodiment of the present invention. Figure 5 is an explanatory diagram showing the optical relationship between the slit light irradiation and the camera, and Figure 6a
-d are explanatory diagrams of arithmetic processing, Fig. 7 is a flowchart showing the arithmetic processing order, Fig. 8 is an explanatory diagram of viewing angles of slit patterns, Fig. 9 is an explanatory diagram of a plurality of circular slit patterns, and Fig. 10 is an explanatory diagram of a plurality of circular slit patterns. Figures a to d are explanatory diagrams of the process for determining the optimal circular slit pattern, Figure 11 is a flowchart of the process for determining the optimal circular slit pattern, and Figure 12 is an explanatory diagram showing the projection characteristics of the lens.
FIG. 13 is an explanatory diagram of a conventional optical system for determining the center of a lens, and FIG. 14 is an explanatory diagram of a screen image in the conventional example. In FIG. 1, 111 is a lens to be measured;
3 is a circular pattern light, and 115 is a light irradiation means. In FIGS. 2 to 13, 21 is a lens;
33, 211 is a lens to be measured, 212 is a television camera, 31, 214 is an autocollimator, 217
is a mirror, 230 is a measurement section, 239 is a calculation section, 251
is a slit light source, 261 is circular image data, 331
is a slit.

Claims (1)

[Scope of Claims] 1. Projection of a lens having a light irradiation means 115 capable of injecting light from a predetermined direction toward a lens of an imaging system, and a measuring means 121 for measuring where on the imaging surface this light is imaged. In the characteristic measuring device, the light irradiation means 115 has a plurality of concentric pattern light irradiation functions 113 that are circular slit parallel lights to the lens to be measured 111 of the imaging system; When determining the center of a circular image on the imaging surface based on received data of concentric circular pattern light, the present invention is characterized by having a function of determining an optimal circular image whose center should be determined from a plurality of received circular images. Projection characteristic measuring device.