WO1988003345A1

WO1988003345A1 - Image analysis system

Info

Publication number: WO1988003345A1
Application number: PCT/GB1987/000761
Authority: WO
Inventors: Charles Thomas Austin
Original assignee: Charles Thomas Austin
Priority date: 1986-10-30
Filing date: 1987-10-28
Publication date: 1988-05-05
Also published as: AU616573B2; KR880702028A; AU4433089A; JPH01501114A; GB8625957D0; GB2197463A; AU595953B2; EP0292514A1; GB2197463B; AU8105387A

Abstract

An image analysis system includes a means (2) for forming an image of an object (1), and an illumination apparatus, the illumination apparatus comprising: illumination means (5), for illuminating the object; and processing means (3), for comparing the brightness of the image with a predetermined brightness level or brightness levels, and capable of increasing or decreasing the light output of the illumination means on the basis of the result of said comparison. Also disclosed is a system including a measurement apparatus for comparing the size of an object with that of a reference object (10), the apparatus comprising: a means (2) for forming an image of the object whose size is to be determined, when the object is in a viewing position; and a reference object, mounted in the apparatus, which may be viewed, using the means for forming an image, instead of, or in addition to, the object whose size is to be measured. The invention is disclosed with particular reference to its incorporation in a static indentation hardness testing machine.

Description

IMAGE ANALYSIS SYSTEM

This invention relates to a system for analysis of an image, such as one produced by a closed-circuit television (CCTV) camera. In preferred embodiments, the invention is applicable to a static indentation hardness testing machine, in which an image of an indentation is analysed using a CCTV camera and a processor.

It is well known that, in order to be able to analyse accurately an image using such a device, it is necessary for the object to have optimum illumination. For many kinds of object this presents.no particular difficulty; however, it is extremely difficult to illuminate metal obj cts, which may have differences in surface finish and reflectivity, sufficiently well to be able to analyse the image satisfactorily. A consequence of this is that it is generally necessary to compromise on the illumination, with the result that many images of metal objects are not accurately analysed.

The present invention seeks to overcome these difficulties, by providing a system in which the amount of illumination of an object may be varied, in order to achieve optimum illumination of the object.

In many applications, for example when using a static indentation hardness testing machine, there is a need to determine the size of an object (such as an indentation in a sample) which is viewed through a CCTV camera, using an image processor. In such applications, it is conventional to calibrate the processor by presenting to the CCTV a reference object, which forms a calibration image, and which might take the form of a calibrated scale, a representation of an object of the type under test, or an actual sample of an object of the type under test, the representation or the sample being of known size. If data concerning the size of the reference object are input into the processor, which is able to generate a measurement of the image size in units such as pixels or chord lengths, the sizes of objects subsequently placed in the field of view of the CCTV camera may be determined. - These measurements may be made by comparison between the size of the object in units such as pixels or chord lengths compared with the known size of the calibration image in those units.

Such a system, however, has major practical disadvantages. In general, the calibration image is placed by hand, and needs to be positioned extremely accurately if subsequent measurements are to be made accurately, while the data concerning the size of the calibration image must be input manually into the processor, with the risk of error by the operator. In addition, whenever re-calibration of the processor is carried out, the measuring operations of the machine must be interrupted. As a result, there is a tendency to re-calibrate.only infrequently. However, images produced by CCTV cameras may drift, or distort, with time or with variations in temperature. Such changes may have a serious effect on the accuracy of a measurement system. This is particularly important because of the time taken to re-calibrate a processor, and hence the infrequency with which re-calibration is carried out.

The present invention seeks to overcome the disadvantages associated with known calibration systems, by providing an image analysis system in which the calibration image is correctly and automatically' positioned, without the need for an operator, and without interruption of the sequence of measurements. Thus, the possibility of operator error, and the effects of camera drift with time and/or temperature, are substantially eliminated because re-calibration can be achieved in a short time, and hence more frequently.

According to a first aspect of the present invention, there is provided an image analysis system including a means for forming an image of an object, and an illumination apparatus, the illumination apparatus comprising: . illumination means, for illuminating the object; and processing means, for comparing the brightness of the image with a predetermined brightness level or brightness levels, and capable of increasing or decreasing the light output of the illumination means on the basis of the result of said comparison.

According to a preferred embodiment of the first aspect of the invention, there is provided a static indentation hardness testing machine, for determining the hardness of a sample on the basis of the. size of an indentation made in the sample, the machine including an image analysis system as defined above.

Preferably, the means for forming an image of the object comprises a closed-circuit television system, although, alternatively, an image array may be used. When the image analysis system is used in conjunction with a static indentation hardness testing machine, the illumination means is arranged to illuminate the surface of the sample surrounding the indentation. Preferably also, the image analysis system comprises a processing means, for determining a grey level of the image at one or several locations in the image, and calculating a mean grey level on the basis of said measured grey levels. In the preferred embodiment, the calculated mean grey level is then compared with two predetermined grey levels forming the end-points of a predetermined range. If the calculated mean grey level is outside the predetermined range, the light output of the illumination means is increased or decreased as appropriate.

According to a second aspect of the present invention, there is provided an image analysis system including a measurement apparatus for comparing the size of an object with that of a reference object, the apparatus comprising:- a means for forming an image of the object whose size is to be determined, when the object is in a viewing position; and a reference object, mounted in the apparatus, which may be viewed, using the means for forming an image, instead of, or in addition to, the object whose size is to be measured.

In preferred embodiments of the second aspect of the invention, the object, the size of which is to be measured, and the reference object may be viewed via a beam splitting mirror, and are located such that an image of one is obtained by transmission of light through the beam splitting mirror and an image of the other is obtained by reflection of light from the beam splitting mirror. Preferably, the object whose size is to be determined is located at the viewing position, which is located in the focal plane of a CCTV camera lens, and illuminated, while the reference object is also located in a focal plane of the camera lens, and is illuminated, and hence viewed through the camera, only when there is no object in the viewing position. For a better understanding of the present invention, and to show how it may be carried into effect, reference will now be made, by way of example, to the accompanying drawings in which:-

"Figure 1 is a block schematic diagram showing an image illumination apparatus, forming part of an image analysis system in accordance with a first aspect of the pres.ent invention;

Figure 2 is a cross-section through a measurement apparatus, forming part of an image analysis system in accordance with a second aspect of the invention; and Figure 3 is a view from below of the measurement apparatus shown in Figure 2.

Figure 1 shows, schematically, an image illumination apparatus according to an embodiment of the first aspect of the invention. A sample 1 is viewed through a television camera 2. In the preferred embodiment of the invention, the sample is a sample from a hardness testing machine, and bears an indentation, the size of which is to be measured. The image from the camera 2 is transferred to a frame store 4 and then to a processor 3, in which the image is analysed. The processor has access to the contents of an associated memory 20, and controls the voltage supplied to a lamp 5 which illuminates the object 1. It has been found that accurate analysis of an image, and hence accurate measurement of the size of the indentation in the sample, can be consistently achieved if the mean grey level at certain points in a fixed pattern which surrounds the stored image of the hardness indentation lies within a particular range of grey levels. The fixed pattern is of importance because of the build up of material in the immediate vicinity of a hardness indentation, caused by the displacement of material during the indentation itself. If the level of illumination is uncontrolled, the mean grey level found at the same points will vary widely, depending upon the reflectivity of the sample in which the indentation is made. The reflectivity will depend upon the nature of the metal itself, and on the surface finish of the sample.

The image viewed by the camera 2 is divided, by the frame store 4, into pixels, to each of which is assigned a grey level between 0 and 63. Appropriate programming of the processor allows the size of the largest anticipated indentation to be identified, on the basis of known data about the hardness test. The processor then selects a number of pixels forming a pattern approximately surrounding that notional indentation and calculates the mean grey level of these pixels. This ensures that the sampled pixels do not, in fact, form part of the image of the actual indentation itself, but only of the surrounding material. As a result of a large number of such tests, it has been possible to determine a range of such mean grey levels for which an accurate analysis of the image may be achieved. Therefore, the calculated mean grey level in each case is compared with the end-points of that range, which are stored in the memory 20. If the calculated mean grey level is below the desired range it is necessary to increase the illumination provided by the lamp 5, while if the calculated mean grey level is above the required range, it is necesary to decrease the level of illumination. * In an alternative embodiment, analysis of the image may be carried out in real time, if the processor 3 has sufficient memory capacity. In this case, it would not be necessary to provide the frame store as a separate device.

The lamp is of a conventional type, and is controlled by a lamp controller which is able to provide any one of 16 available discrete voltage outputs. Therefore, if it is required to increase the level of illumination, the voltage to the lamp is increased by one step, while if it is required to decrease the level of illumination, the voltage to the lamp is decreased by one step. In order to obtain a mean grey level within the desired range, it may be necessary to repeat this operation,, and to alter the voltage supplied to the lamp several times.

Figures 2 and 3 illustrate the apparatus used for mounting the object whose size is to be measured and the reference object, in order to allow the reference object to be viewed without the need for interrupting the measurement operations, and to substantially eliminate measurement errors caused by camera drift.

A camera extension tube 6 is mounted horizontally to the front of the camera 2 shown in Figure 1 , and is mounted inside a support tube 7. At the end of the camera extension tube 6 is located a beam splitting mirror 8, behind which is provided a front silvered mirror 9. Above the front silvered mirror is positioned a reference object 10, which is back-lit by a light source mounted in a holder 11. The beam splitting mirror is mounted in optical blocks 12, 13, which are connected to further optical blocks 14, 15. The optical blocks are provided with polarisers 16, 17, and, adjacent the beam splitting mirror, there is provided an optical window 18, and, on the other side of the mirror, a diffuser 19.

This apparatus is intended to be used in connection with a hardness testing machine, and in particular a static indentation hardness testing machine for performing, for example, the Brinell hardness test. In this particular test, a measure of the hardness of a material is formed by measuring the size of an indentation formed in a sample as a result of the application of a known load. The mean diameter of the indentation formed^"by a spherical ball is inversely proportional, for a given load, to the Brinell Hardness Number of the material. A sample, with an indentation formed therein, is positioned below the optical wndow 18 along the axis B. An image of the indentation is then reflected by the beam splitting mirror along axis A to the camera via the extension tube 6, which contains a lens to enlarge the image sufficiently for measurement purposes. The beam splitting mirror 8 is chosen such that approximately 50% of incident light is transmitted through the mirror and approximately 50% is reflected. Thus, it is possible to illuminate the indentation from above the mirror, and this is done by positioning a light source, such as a fibre optic cable remotely supplied with light from the lamp 5, along the axis C. As a result of the properties of the beam splitting mirror, it is also possible to view, through the camera, a second image. In this embodiment of the invention, an image of a back-lit reference object 10 is reflected from the front silvered mirror 9, and partially transmitted through the beam splitting mirror 8 along the extension tube 6 to the camera 2. The reference object is an accurately made representation of a hardness indentation, surface etched onto, polished opal glass. This object is mounted in the focal plane of the camera, and is visible to the camera only when it is back illuminated from a light source mounted in the holder 11 along the axis D.

As^" the reference object is fixed in position, it is illuminated only when it is required to re-calibrate the apparatus. Thus, the back illumination of the reference object is used only intermittently, while, for convenience, the lamp 5, illuminating the sample, is left on permanently. As it is required to view only one of the images at any time, it is necessary to prevent stray light from the light source along the axis C illuminating the reference object 10 via the mirrors 8,9. This is achieved by means of the cross polarised filters 16, 17.

When first used, it is necessary initially to calibrate the apparatus. The reference object is back- lit, and no hardness indentation sample is present along axis B. Therefore, the only image visible is that of the reference object, which is of known size. Therefore, the image analysis system calibrates itself by means of the processor, by measuring the number of pixels occupied by the calibration image, and determining a^Λrelationship between the size of the object in millimetres and the measured number of pixels. Then, the light at D is switched off, and a metallic sample bearing a hardness indentation is moved into the focus of the camera along axis B. The camera now records the image of this hardness indentation, and, after making any necessary adjustments to the illumination of the sample, applies the relationship between pixels and millimetres derived from the calibration procedure to the size of the image, in pixels; thus, a measure of the size of the hardness indentation is achieved. Using this system, it is possible for the apparatus to be regularly re-- calibrated, since the time taken for the calibration is very short, and it ^'may be carried out while one hardness indentation is being removed from the appratus and another is being introduced. ^"* *

Thus, there is provided an image analysis system, which may be rapidly and accurately calibrated, in order to substantially eliminate the possibility of errors caused by camera drift or operator error. In addition, the invention provides a system which allows accurate analysis of the image to be made, by ensuring correct illumination of the object.

Claims

1. An image analysis system including a means for forming an image of an object, and an illumination apparatus, the illumination apparatus comprising: illumination means, for illuminating the object; and processing means, for comparing the brightness of the image with a predetermined brightness level or brightness levels, and capable of increasing or decreasing the light output of the illumination means on the basis of the result of said comparison.

2. A system as claimed in claim 1, wherein the means for forming an image of the object comprises a closed-circuit television system.

3. A system as claimed in claim 1 , wherein the means for forming an image of the object comprises an image array.

4. A static indentation hardness testing machine, for determining the hardness of a sample on the basis of the size of an indentation made in the sample, the machine including an image analysis system as claimed in any preceding claim.

5. A static indentation hardness testing machine as claimed in claim 4, wherein the processing means is adapted to determine the ^'brightness of the image from the brightness of the surface of the sample surrounding the indentation.

6. A static indentation hardness testing machine as claimed in claim 4 or 5, wherein the processing means is adapted to determine a grey level of the image at several locations in the image, and to calculate a mean grey level on the basis of said measured grey levels.

7. A static indentation hardness testing machine as claimed in claim 6, wherein the calculated mean grey level is compared with two predetermined grey levels forming the end-points of a predetermined range, and, if the calculated mean grey level is outside the predetermined range, the light output of the illumination means is increased or decreased as appropriate.

8. An image analysis system including a measurement apparatus for comparing the size of an object with that of a reference object, the apparatus comprising:- a means for forming an image of the object whose size is to be determined, when the object is in a viewing position; and a reference object, mounted in the apparatus, which may be viewed, using the means for forming an image, instead of, or in addition to, the object whose size is to be measured.

9. A system as claimed in claim 8, wherein the object, the size of which is to be measured,^" and the reference object may be viewed via a beam splitting mirror, and are located such that an image of one is obtained by transmission of light through the beam splitting mirror and an image of the other is obtained by reflection of light from the beam splitting mirror.

10. A system as claimed in claim 8 or 9, wherein the object whose size is to be determined is located at the viewing position, which is located in a focal plane of a camera lens, and illuminated, while the reference object is also located in a focal plane of the camera lens, and is illuminated, and hence viewed through the camera, only when there is no object in the viewing position.