WO2005114567A1

WO2005114567A1 - A method and apparatus for measuring a three dimensional object

Info

Publication number: WO2005114567A1
Application number: PCT/GB2005/001401
Authority: WO
Inventors: John Douglas Rogers
Original assignee: Rolls-Royce Plc
Priority date: 2004-05-21
Filing date: 2005-04-12
Publication date: 2005-12-01
Also published as: GB0411469D0

Abstract

Apparatus and a method for determining the three dimensional positions of features (A1, A2) of an object (A). The apparatus comprising a camera (16) connected to image processing means (12). The camera (16) is mounted upon a means (14) for moving the camera (16) around a known fixed symmetric path (16) within a plane substantially perpendicular to the camera axis (6). Preferably the known fixed symmetric path (18) is a circle. Means (20, 22, 24) are provided to maintain the orientation of the camera (16) as it is moved around the symmetric path (18). The method comprising calibrating the apparatus (10), moving the camera (16) about a fixed symmetric path (18) within a camera plane, and from a series of positions around the symmetric camera path (18) capturing a series of images of the object (A) Then a feature of interest (A1, A2) of the object (A) from the images is identified from the images. The images of the features (A1, A2) from the series of images are then tracked and an apparent image path (34, 36) of the identified feature of interest (A1, A2) of the object (A) is determined. The apparent shape and dimensions of the image path (34,36) of each identified feature (A1, A2) is then analysed against that of the known camera path (18) to determine characteristic features of the image path (34, 36) of that feature, and determine the position of the feature (A1, A2) using the characteristic features (ΦA1ΦΦA2) of the image path (34, 36) and the calibration.

Description

A METHOD AND APPARATUS FOR MEASURING A THREE DIMENSIONAL OBJECT

The present invention relates generally to method and apparatus for measuring the three dimensional positions of features of an object. Optical three dimensional measurement techniques are used to optically determine the three dimensional coordinates of features of interest in a number of imaging techniques including, for example X-ray radiography, manufacturing NDE visual inspection, engine inspection, hologram visualisation, three dimensional microscopy. Typically such three dimensional optical measurement techniques are based upon analysis of stereo image pairs to derive the three co-ordinate positions of the features. The stereo image pairs are obtained either using two cameras set up on a stereo base line, or alternatively a single camera which is displaced to obtain separate images of the object can be used. In both cases images as would be seen from a left' and right' view are obtained and analysed to provide the position of features of the objects viewed. The analysis required is complex. The two camera system requires the introduction of correction factors to correct the different image distortion effects of the different cameras, and for the single camera system requires accurate control of the movement of the camera, although the calculations are simplified because the same camera and image distortion is used. Furthermore to obtain accurate position information multiple stereo pairs aligned along a number of axes need to be taken. Examples of such previous techniques are described in ^Calibrating a Mobile Camera' by P Puget and T Skordas, Image and Vision Computing, Volume 8, Number 4, November 1990; and in Computation of a cross-section structure: a projection-based approach, by Myoung J Kim and Joon H Han, Image and Vision Computing 16 (1998) 653-667. It is therefore desirable to provide a simplified and improved method and apparatus for measuring a three- dimensional object that addresses these problems and/or offers improvements generally. According to a first aspect of the present invention there is provided apparatus for determining the spatial position of a feature of an object comprising a camera connected to image processing means; characterised in that means for relative movement between the camera and an image of the object is provided, the means for relative movement defines a known path. Preferably, the spatial position of a feature is in three dimensions. Preferably, the known path is substantially defined within a plane substantially perpendicular to an axis of the camera. Preferably, the known path is circular. Preferably, the path is symmetrical about a mutually perpendicular axis relative to an axis of the camera. Preferably, the camera is moved relative to a stationary object or image of the object. Alternatively, the object is moved relative to a stationary camera or the image of the object is moved relative to a stationary camera. Preferably, the means for relative movement of the camera comprises an arm, the camera or object mounted upon a first end of the arm with the other end of the arm rotatably mounted so that the arm and camera or object mounted thereon can be rotated about a fixed axis passing through a second end of the arm. Preferably, the means are provided to maintain the orientation of the camera or object or image of the object as it is moved around the path. Preferably, the means to maintain the orientation comprise a fixed pulley mounted coaxially with the axis of rotation of the arm and a free pulley upon which the camera or object is mounted, the free pulley rotatably mounted on the first distil end of the arm to rotate about an axis substantially parallel to the axis of rotation of the arm, a linkage means provided between the pulleys to link the rotation of the pulleys. Alternatively, the means for relative movement comprises a track capable of mounting any one of the group comprising the camera, the image of the object or the object. According to a second aspect of the present invention there is provided a method 13. A method of determining the spatial position of a feature of an object using a camera and image processing apparatus, the method comprising the steps of a) calibrating the apparatus, b) moving the camera or object or an image of the camera about a fixed path within a camera plane, c) from a series of positions around the camera or object or image path capturing a series of images of the object, d) identifying a feature of interest of the object from the images, e) tracking each feature of interest from the series of images and determining an apparent image path of the identified feature of interest of the object, f) analysing the apparent shape and dimensions of the image path of each identified feature with that of the known camera path, to determine characteristic features of the image path of that feature, g) determining the position of the feature using the characteristic features of the image path and the calibration. Preferably, the fixed path about which the camera or object or image of the object is moved is circular. Preferably, the apparent path of the image path is a circle, and the diameter of the circle is used to determine the position of the feature of interest is from the camera. Preferably, the centre of the image path circle represents the position of the feature of interest in the plane perpendicular to the camera axis and is use to determine said position of the feature of interest in said plane. Preferably, a best fit circle is fitted through the apparent path of the image of the feature of interest. Preferably, the feature of interest of the object from the images is automatically determined and correspondingly identified in the series of images by an image recognition system. Preferably, the camera, object or image of the object, the camera having an axis, is moved about a fixed symmetric path within a camera plane with the axis of the camera maintained substantially perpendicular to the camera plane within which the camera, object or image of the object is moved . Preferably, the camera axis is maintained substantially parallel to an axis of the symmetric path perpendicular to the camera plane. Preferably, the camera has a camera axis which is arranged to pass through a fixed point as the camera, object or image of the object is moved around the path. The present invention will now be described by way of example only with reference to the following figures in which: Figure 1 is a schematic illustration of an embodiment of the apparatus for optically measuring the three dimensional positions of features of an object according to the present invention; Figure 2 is a more detailed side view of the apparatus shown in figure 1; Figure 3 is a simplified schematic representation showing the imaging of two points following movement of the camera; Figure 4 is a simplified schematic illustration showing the imaging of two points using a camera; Figure 5 is a simplified illustration of the images of two points on an object produced by the apparatus of figure 1 operating in accordance with the method of the present invention. Figures 6A and 6B are a simplified schematic representation of an alternative optical system using mirror to move an image of the object relative to the camera. By way of an exemplary embodiment, a single camera 16, connected to suitable image processing equipment 12, is arranged to trace out a complete symmetrical path 18 of fixed dimensions (for example a circular path) . The field of view of the camera 16 is arranged to enclose the volume of interest of the three dimensional object A. Points of interest, in this case Al, A2, which are required to be measured, are identified on the object A. The camera 16 has a fixed orientation relative to a central axis 2 of the circular path 18 and takes images of the object A from a number of positions along the circular path 18. If these images are superimposed then the image of a single point Al, A2 of interest will also trace out a circle 34, 36 (See figure 5) . The diameter φ_Aι_>- φ_A2 of this image circle 34, 36 is a function of the diameter D of the circular camera path 18 and the distance zl, z2 (z axis position) of the point Al, A2 of interest from the camera 16. Since the diameter D of the circular camera path 18 is fixed then the distance ZAI/ z_A2 of the point of interest Al, A2 from the camera 16, and their relative positions, can easily be calculated from the diameter φ_al, φ_A2 of the image circle 34, 36. The position x_Ai, x_A2, y_Aι, y_A2 of the point of interest Al, A2 relative to the origin 0, and central axis 2 is given by the position of the centre 46, 48 of the image circles 34, 36. Alternatively these positions x_A1, x_A2, y_A1, y_A2 can be calculated from any single image taken from around the circular camera path 18 provided the relative position of the camera 16 around the circular path 18 is known. In more detail, referring to figure 1 the apparatus 10 for measuring the three dimensional positions of features Al, A2 of an object A comprises a single camera 16 which is connected 8 to an image processing system 12. The image processing system 12 typically comprises a conventional computer programmed with suitable image processing and manipulation software. The camera 16 is mounted upon a camera support mechanism 14 and is arranged with the field of view of the camera 16 arranged to enclose the volume of interest of the three dimensional object A. The camera support mechanism 14 securely supports the camera 16 and is arranged to move the camera 16 around a complete fixed circular path 18 about a central axis 2. As the camera 16 is moved around the path 18 the mechanism 14 ensures that the orientation of the camera 16 is maintained and that the viewing axis 6 of the camera 16 is parallel to the central axis 2. The camera mounting arrangement 14 is shown in more detail in figure 2. The mounting arrangement 14 comprises a fixed stand 28. A radial arm 26 is pivotally mounted at, and extends from, the distil end of the stand 28 so that the arm 26 rotates about the central axis 2. A motor drive 30 mounted upon the stand 28 is arranged to rotate the radial arm 26 via suitably arranged gears 31, 32. The camera 16 is pivotally mounted, about a camera mounting axis 4, on the distil end of the radial arm 26. The camera mounting axis 4 is parallel to the central axis 2 and the camera 16 is mounted so that the camera axis 6 and camera mounting axis 4 are coincident and/or parallel. To maintain the orientation of the camera 16 a pulley 22 disposed about the camera mounting axis 4 is attached to the camera 16. Attached to the stand 28 there is a corresponding fixed pulley 20 mounted about the central axis 2 about which radial arm 26 rotates. A belt 24 interconnects the pulleys 20, 22. Consequently as the arm 26 rotates, pulley 22 will be rotated relative to the arm 26 by the belt 24 by an angle sufficient to maintain the original orientation of the camera 16 around the path 18 and to keep the camera 16 axis 6 parallel to the central axis 2. It will be appreciated however that other mechanisms and methods could be used in order to move the camera 16 around a fixed path 18 and to maintain the camera 16 orientation. The length of the radial arm 26 is also preferably adjustable so that the diameter D of the circular path 18 can be adjusted to suit the required field of view required for the camera 16 to view the object A or features Al, A2 thereof from all positions around the circular path 18. The analysis method and principle behind the method used by the system to determine the three dimensional positions of points of an object A is illustrated with reference to figures 3 and 4. Referring to figure 3, consider two points PI and P2. Point Pi is in a plane 11₁ perpendicular to the central axis 2 and camera axis 6 and at a distance 1 in the z direction from the origin 0 and camera 16. Point P2 is aligned in the z direction with point PI and is contained within a plane π₂ a distance d from the first point PI. With the camera 16 aligned with the points Pi, P2 so that the camera axis 6' passes through the points PI, P2, the images Pi', P₂' of points PI and P2 in the camera image plane π_c will lie at the origin 44 and centre of the camera image plane lie. Moving the camera a distance e in the x direction as shown by arrow 40 means that the point PI and P2 are now a distance e from the translated camera axis 6' ' . Points PI and P2 will now be focussed by the camera lens 42 to points Pi'' and P₂' ' in the camera image plane lie. In the camera image plane Etc point Pi will appear to move from point PI' to point Pi'' as the camera 16 is moved. Similarly point PI will appear to move from point P2' to point P2' ' . The origin 44 of the camera image plane lie is translated with the camera axis 6. Therefore in the camera plane lie point Pi will appear to move a distance v in the x direction and point P2 will appear to move a distance u, as the camera is translated. Analysing the lengths shown in figure 3 gives by comparing similar triangles :- e v For point Pi 1 ^~ f which gives 1=≥ eqn. V e u and for point P2 1+d f which gives l+d =≥ eqn. u

where f is the focal length of the camera 16.

Subtracting equation 1 from equation 2 gives :-

which gives :- fe = eqn. 3 v-u

In the system 10 the camera 16 movement e will be constant and fixed for all the measurements. Similarly the focal length f of the camera 16 is constant for all measurements using that camera 16. Therefore dvu fe= = K eqn. 3a u-

where K is a constant for that particular camera 16 and camera movement e . If the system is calibrated using two points PI, P2 separated by a known fixed distance d in the z direction and the apparent perceived movement u, v of the image points

PI' ' , PI' ' is recorded then the constant K for the system 10 can be calculated. An equation similar to equation 1 is valid for any point P a distance z from the camera. Consequently the z distance from the camera 16 is given by:-

K z=— eqn. 4 w where z is the distance from the camera 16 in the z direction, K is the calibrated constant calculated as described above and w is the measured movement of the image of the point P recorded by the camera 16 as it is moved a known distance e. The above analysis is for a camera 16 moving a fixed distance e in the x direction. The analysis can however be extended by analogy to a camera 16 moving in a circular path

18. In such a case the image points will now move in a circular path as the camera 16 is also moved in a circular path 18, with the distances u, v, w moved by the image points (PI' to PI", P2' to P2" , and P' to P" ) equating to the diameter Φ of a circular path form by the image points. This gives :- K' z = — eqn. 5 Φ where z is the distance of the point P from the camera in the z direction, K' is the calibration constant for the system 10 moving in a circular path 18 calculated in a similar way to K but using measurements taken from the camera system 10 moving in a circular path 18, and Φ is the diameter of the circle mapped out by the image points of a point P recorded by the camera 16. Therefore the system 10 is calibrated using two points PI, P2 aligned in the z direction and separated by a known distance d. A value for the constant K' can then be determined from the measurement of the diameters Φ of the circular image paths through which the images of the points PI and P2 move using equations 3 and 3a. Once this is done and the system 10 accordingly calibrated, it can then be used to determine the z position of any other point P from the measurement of the diameter Φ of the image path through which the images of that point move using equation 5. An example of the circular paths 34,36 produced is shown in figure 5 for two points Al, A2 on an object A. As the camera moves around the circular path 18 the images 38, 40 of the points Al, A2, taken from positions around the circular path 18, trace out a circle 34, 36 of respective diameters Φ_Aι, Φ_A2. Using these measured diameters Φ_Aι, Φ_A2 the distance z in the z direction from the camera 16 can be determined. The principle for determination of the x, y coordinates is explained with reference to figure 4. Figure 4 shows how two points Ql and Q2, aligned in the z direction, are imaged by the camera 16 and camera lens 42, with a focal length f, to image points Ql' , Q2' in the camera image plane Lie. Point Ql is at a distance 1 from the camera 16 in the z direction and a distance t from the camera axis 6 in the x direction. Point Q2 is at a distance d from Ql and is also a distance t from the camera axis 6. Point Ql is focussed by the camera lens 42 to a point Ql' a distance s from the origin 44 of the camera plane lie in the x direction. Similarly point Q2 has an image Q2' in the camera image plane lie a distance r in the x direction from the origin of the camera image plane l e. Again analysing these lengths and comparing similar triangles gives :- t s for point Ql —=— I f tf which gives 1=— eqn. s t r for point Q2 1 + d tf which gives l+d = — eqn. 7 r as before subtracting equation 6 from equation 7 gives: tf tf d =

, . i • _J= ^df rs ^ which gives f ^{= —} eqn. t s -r)

Again if the system is calibrated using two points Ql, Q2 separated by a known distance d, a known distance t from the camera axis 6, and the position of the image points Ql' Q2' in the x direction is measured, then the value f for the camera can be calculated using equation 8. This value f will then be a constant for the camera system.

rearranging equation 6 gives:- f applying this more generally for any point Q a distance z from the camera 16 in the z direction, and a distance x from the camera in the x direction gives :- zs x = —— eqn. 9 f where s_x is the measured distance of the image point Q' in the camera image plane Tic from the origin of the camera image plane Tic. A similar analysis can be carried out in the orthogonal y direction giving: - zs y = —— eqn. 10 f

Therefore using equations 9 and 10 the x and y positions relative to the camera axis 6 can be determined once the system has been calibrated for f and the z distance has been calculated using equation 5 as described above. Applying this analysis to a camera system 10 that moves in a circular path 18. Figure 5 shows the image circles 34, 36 produced of points Al, A2 as the camera 16 is moved around a circular path 18. The centres 46,48 of the image circles 34,36 correspond to the x and y positions of the images of the points Al, A2 that would be obtained if the camera 16 were mounted at the centre 0 of the circular path 18 around which the camera is moved. Therefore using the measured values s_xAι, s_xA2, s_yAι, s_yA2 of the central image points 46, 48 the x, y positions of the points Al, A2 relative to the centre of the camera movement path 18 can be determined using equations 9 and 10. The z positions of the points Al, A2 needed in equation 9 and 10 would have been determined already from equation 5 above. It will be appreciated that although in theory the images of the points of interest Al, A2 recorded by the camera 16 should trace out a circle, in practice this may not occur. In particular the camera 16 and camera lens 42 is likely not to be perfect. The image processing system 12 therefore carries out standard conventional processing to fit a circle through the recorded image points 34, 36 of points Al and A2 the object A. A number of conventional circle fitting analysis systems are known in the art and any such system could be used. For example a circle could be fitted to the points using the Hough transform technique. In addition, the camera 16 and lens 42 may introduce distortion effects into the image, with a result that an ellipse or similar shape may be produced instead of a circle. Similarly, although preferably perpendicular, the imaging plane does not have to be perpendicular or in a single plane. To overcome these distortion effects the images recorded by the camera 16 may be corrected before they are analysed. Such correction for camera 16 and lens 42 distortion is relatively standard in the art of imaging. The correction may be carried out by first calibrating the camera 16 by taking an image of a square test grid with its plane normal to the camera axis 6. The image of the test grid should also be of a grid form. Accordingly from the image of the test grid each pixel in the camera 16 image is given a correction factor so that the image of the test grid equates to the actual test grid. These correction factors can then be applied to the other images taken by the camera 16. The images recorded in the camera image plane lie by the camera 16 comprise a rectangular array of individual identical pixels. The measurement in the camera image plane lie of the various image points may therefore conveniently be measured in pixel units. In applying the system in practice to determine the position of points of interest of an object A, the points of interest Al, A2 must be identified from the image of the object A. Furthermore the images of the distinct points of interest Al, A2 must be tracked through the series of images that are taken as the camera 16 is moved around the path 18 in order to generate the image circles 34, 36 for the respective points Al, A2. Accordingly before the analysis to determine the x, y, z positions of the points of interest Al, A2 can be carried, as described above, the image processing system 12 first identifies the points of interest Al, A2 from the images recorded by the camera 16 as it moves around the path 18. The image processing system 12 then tracks the movements of each of these points within the camera image plane Lie through the successive images from around the path 18. Suitable image processing techniques to carry out this image point identification and tracking, which are embodied within the image processing system 12, are well known in the art. Such systems include for example edge detection and automatic template tracking systems. The analysis and determination of the x, y, z positions (as described above) is then carried out on the tracked image paths and points. Overall the method comprises the following steps :- a) Calibrate the system 10 using two points aligned in the z direction and separated a known distance d apart. From the calibration calculate the constants f and k' for the system 10. b) Move the camera 16, object or image of the object around a fixed circular path 18 and record a series of images of the object A from various positions around the path 18. c) Analyse each of the images taken of the object A from around the camera path 18 to identify the desired points of interest Al, A2 in each of the images taken. d) For each of the desired points of interest Al, A2, plot the positions 38 within the camera image plane lie of the desired point of interest Al recorded from each of the images taken. e) Using circle fitting techniques fit an image circle 34 through the points 38 plotted in d above. f) Determine the diameter Φ_Aι and centre 46 of the image circle 34 plotted in e above. g) Determine the z position of the desired point of interest Al from the camera 16, using equation 5 and the diameter Φ_Aι of the image circle 34 obtained from f. h) Determine the x position x_Aι of the desired point of interest Al from the x position s_xAi of the centre 46 of the image circle 34 using equation 9. i) Determine the y position y_Aι of the desired point of interest Al from the y position s_yAi of the centre 46 of the image circle 34 using equation 10. j) Repeat steps d to i for each of the other desired points of interest A2. An advantage of this method is that the analysis process to determine the x, y, z positions x_A1, x_A2, y_A1, y_A2 of the points of interest Al, A2 is considerably simplified as compared to the conventional optical stereo analysis process. The analysis is reduced to fitting a circle 34,36 to the superimposed/plotted images of the points produced and calculating the diameter Φ_Ai, Φ_A2 of that circle 34, 36. In addition, if from some positions on the circular path 18 a feature of interest Al, A2 is obscured by another part of the object A the analysis can still be carried out since the circle fitting and calculation of the diameter Φ_Aι, Φ_A2 does not require a full circle to be defined by the images. Since the images are taken from a number of points around the circular path 18, and so from a number of different angles, accurate position information is obtained using this method and apparatus. The exact movement, and position of the camera 16 around a circular path 18 is also not important to the measurements. This makes the control system for the camera 16 movement simpler with the camera 16 only being required to execute a circular path 18 of a fixed diameter D. The invention has been described with reference to moving the camera 16 around a circular path 18, and this is clearly the preferred method since circle fitting and measuring the diameter of a circle is relatively simple. In addition since a circle has a constant dimension through the diameter the calculations and calibration is simplified. It will be appreciated though that the method and its principles, could equally be applied to moving the camera 16 around a non-circular fixed symmetrical path. In such a case the image paths will similarly correspond to the camera path and accordingly the image paths can be used in a similar way to determine the x, y, z positions of a feature of interest. The characteristic features of the shapes of image paths produced by movement about such a different path will then correlate to the x, y, z positions of the features on the object, in a similar manner to the way in which the diameter and centres of the image path circles were used for a circular movement. It is preferred to keep the camera axis 6 perpendicular to the camera plane Lie and substantially parallel to the central axis 2. However in other embodiments the camera 16 could be arranged and mounted so that the camera axis 6 is directed to and passes through a fixed point on the central axis 2 some distance from the camera plane lie. As the camera 16 is moved around the path 18 the camera 16 is angled keep the camera 16 orientated, at all times, so that the camera axis 6 passes through the fixed point. With such an arrangement the camera field of viewed is more tightly focussed onto the object A and so a camera with a smaller field of view, and hence reduced distortion, can be used to view the object A. The same principles apply to the analysis to obtain the positional information of features Al, A2 on the object A. The images of point of interest Al, A2 of the object A will still move as the camera is moved around the path 18. The distance that the image points move, and the diameter of image circles 34, 36, is still also dependent on the distance of the point of interest Al, A2 from the camera 16. The analysis and trigonometric calculations are however slightly more complicated. Also any feature at the fixed point at which the camera axis 6 is directed will not appear move in the camera image plane lie as the camera 16 is moved around the path 18. Preferably, although not exclusively, the fixed point though which the camera axis 6 is directed is carefully chosen so as not to correspond to a feature of interest Al, A2. Although the present invention has been described with reference to apparatus in which a camera is moved around a known fixed symmetric path, it will be appreciated that under certain circumstances, it may be desirable to maintain the camera in a fixed position and move the object or the image of the object around the fixed symmetric or known path. Referring now to Figure 6A, an alternative arrangement for providing at least two images of the object to the camera, comprises an optical system including a pair of mirrors 60, 61, each rotatable about their own axes 62, 63. The mirrors 60, 61 are capable of being rotated to provide at least two discrete images of the fixed object to the fixed camera 16. Complimentary rotated positions for the two mirrors 60, 61 are denoted by A and A', B and B' and C and C respectively. A means for pivoting the mirrors 60, 61 is provided at 67. Processing of the image and calculation of the position of an object or point on an object is made as defined hereinbefore with due consideration of the distances 1 and d as described with reference to figures 3 and 4. Referring now to Figure 6B, a further alternative arrangement for providing at least two images of the object to the camera, comprises an optical system including a pair of mirrors 64, 65, each rotatable about their common axes 66. The mirrors 64, 65 may be movably mounted on a similar apparatus as described with reference to figure 2, however, there is no requirement for the mirrors 64 or 65 to maintain their relative orientation and as the mirrors move around the axis 66 they should project the image of the object toward the rotatable mirror 65. Alternatively, the mirrors 64 or 65 may be mounted on a track. The mirrors 64, 65 may also be pivotable via pivoting means 67. The present invention is also realised by interposing prisms for the mirrors and the skilled would immediately arrive at an appropriate arrangement, wherein at least one prism is moved to provide two images of the object to the camera . The skilled artisan will readily appreciate that the present invention may be applied, but not restricted to such diverse fields as X-Ray and neutron radiography, manufacturing non-destructive inspection processes, plant and engine inspections, hologram visualisation inspection and measurement and 3D microscopy.

Claims

1. Apparatus for determining the spatial position of a feature of an object comprising a camera connected to image processing means; characterised in that means for relative movement between the camera and an image of the object is provided, the means for relative movement defines a known path.

2. Apparatus as claimed in claim 1 wherein the spatial position of a feature is in three dimensions.

3. Apparatus as claimed in any one of claims 1 or 2 wherein the known path is substantially defined within a plane substantially perpendicular to an axis of the camera.

4. Apparatus as claimed in any one of claims 1-3 in which the known path is circular.

5. Apparatus as claimed in any one of claims 1 to 4 wherein the path is symmetrical about a mutually perpendicular axis relative to an axis of the camera.

6. Apparatus as claimed in any one of claims 1 to 5 wherein the camera is moved relative to a stationary object or image of the object.

7. Apparatus as claimed in any one of claims 1 to 5 wherein the object is moved relative to a stationary camera.

8. Apparatus as claimed in any one of claims 1 to 5 wherein the image of the object is moved relative to a stationary camera .

9. Apparatus as claimed in any one of claims 6-8 wherein the means for relative movement of the camera comprises an arm, the camera or object mounted upon a first end of the arm with the other end of the arm rotatably mounted so that the arm and camera or object mounted thereon can be rotated about a fixed axis passing through a second end of the arm.

10. Apparatus as claimed in claim 9 in which the means are provided to maintain the orientation of the camera or object or image of the object as it is moved around the path.

11. Apparatus as claimed in claim 10 in which the means to maintain the orientation comprise a fixed pulley mounted coaxially with the axis of rotation of the arm and a free pulley upon which the camera or object is mounted, the free pulley rotatably mounted on the first distil end of the arm to rotate about an axis substantially parallel to the axis of rotation of the arm, a linkage means provided between the pulleys to link the rotation of the pulleys.

12. Apparatus as claimed in any one of claims 1 to 8 wherein the means for relative movement comprises a track capable of mounting any one of the group comprising the camera, the image of the object or the object.

13. A method of determining the spatial position of a feature of an object using a camera and image processing apparatus, the method comprising the steps of:- a) calibrating the apparatus, b) moving the camera or object or an image of the camera about a fixed path within a camera plane, c) from a series of positions around the camera or object or image path capturing a series of images of the object, d) identifying a feature of interest of the object from the images, e) tracking each feature of interest from the series of images and determining an apparent image path of the identified feature of interest of the object, f) analysing the apparent shape and dimensions of the image path of each identified feature with that of the known camera path, to determine characteristic features of the image path of that feature, g) determining the position of the feature using the characteristic features of the image path and the calibration.

14. A method as claimed in claim 13 in which the fixed path about which the camera or object or image of the object is moved is circular.

15. A method as claimed in claim 14 in which the apparent path of the image path is a circle, and the diameter of the circle is used to determine the position of the feature of interest is from the camera.

16. A method as claimed in any one of claims 14 to 15 in which the centre of the image path circle represents the position of the feature of interest in the plane perpendicular to the camera axis and is use to determine said position of the feature of interest in said plane.

17. A method as claimed in any one of claims 13-16 in which a best fit circle is fitted through the apparent path of the image of the feature of interest.

18. A method as claimed in any one of claims 13-17 in which the feature of interest of the object from the images is automatically determined and correspondingly identified in the series of images by an image recognition system.

19. A method as claimed in any one of claims 13-18 in which the camera, object or image of the object, the camera having an axis, is moved about a fixed symmetric path within a camera plane with the axis of the camera maintained substantially perpendicular to the camera plane within which the camera, object or image of the object is moved.

20. A method as claimed in claim 19 in which the camera axis is maintained substantially parallel to an axis of the symmetric path perpendicular to the camera plane.

21. A method as claimed in any one of claims 13 to 18 in which the camera has a camera axis which is arranged to pass through a fixed point as the camera, object or image of the object is moved around the path.

22. An apparatus as hereinbefore described with reference to figures 1 to 6.

23. A method as hereinbefore described with reference to figures 1 to 6.