KR20080100165A - Machine Vision System for 3D Measurement and Inspection in Semiconductor Industry - Google Patents

Abstract

The system 10 according to a preferred embodiment of the invention comprises a structural subsystem 20 for providing a stable platform and vibration isolation, a camera for capturing an image of the moving object as the object moves into the viewing area. Subsystem 30, an illumination subsystem 50 for illuminating a moving object, and a controller for selectively manipulating the illumination subsystem 50 to observe the reflectance of the composite under different conditions. The system 10 is specifically designed to illuminate and capture images of moving objects, such as connecting terminals (eg, lead, ball grids, and pads) of electrical components packaged in manufacturing facilities. The system 10 may, however, also be used to illuminate a suitable object in a moving or stationary state and to capture its image.

Description

MACHINE VISION SYSTEM FOR THREE-DIMENSIONAL METROLOGY AND INSPECTION IN THE SEMICONDUCTOR INDUSTRY

The present invention is a counting-in-part of "Lighting Subsystem for a Machine Vision System," filed June 13, 2006 in the United States. This application claims the (priority) benefit of US provisional application No. 60 / 735,778, entitled “Semiconductor Industrial 3D Measurement and Inspection”.

The present invention relates generally to the field of machine vision and, more particularly, to a new and useful machine vision system for three-dimensional metrology and inspection in the semiconductor industry in the field of machine vision.

In 1976, Yakimovsky and Cunningham wrote an article entitled "A system for Extracting Three-Dimensional Measurement from a Stereo Pair of TV Cameras." The article is published in "Computer Graphics and Image Processing 7, 195-210 (1978)", which is incorporated herein in its entirety. Through a complex series of modeling, calculations, and calibrations, the article discloses accurate position calculations for each camera, which facilitates direct triangulation of the three-dimensional position of the object. Let's do it.

US Pat. No. 6,101,455, incorporated herein, was issued on August 8, 2000 to Michael S. Davis, one of the teachers of one of the inventors of the present invention. The 455 patent refers to Yakimovsky and Cunningham articles and discloses automatic measurement of cameras in the system. The 455 patent includes a process of calibrating cameras to a common coordinate frame and moving an object by a known relative distance to determine the positions of the cameras using a vector relationship. Discuss the technique. The 455 patent does not, however, disclose the use of two or more cameras.

US Pat. No. 6,064,756, incorporated by this invention, was registered on May 16, 200 and is currently owned by Scanner Technologies Corporation. The 756 patent discloses a three-dimensional inspection apparatus for ball array devices and discloses using three cameras. The 756 patent, however, discloses the use of triangulation to calculate the three-dimensional position of a ball in a ball arranging device, based on a pre-calculated calibration plane, and the exact position of the cock camera. In addition, direct triangulation with respect to the three-dimensional position of the ball in the ball arranging apparatus is not disclosed.

Prior art in the art (of the present invention) is suitable for a particular application, while the body and connecting terminals (leads, ball grids, bumps, pads) of the fast moving semiconductor Or for accurate measurement of packaged electrical components in various manufacturing processes. The prior art is simply not suitable for inspection under limited circumstances, such as the deformation of connecting terminals in semiconductor or packaged electrical components. Under these circumstances, errors are caused by statistical errors (randomly generated and thus implicitly unpredictable, fluctuation in the measuring device) and systematic errors (unknown but not random). Caused by fluctuations).

Thus, there is a need to create improved machine vision systems that reduce statistical and systematic errors in the machine vision field, making them suitable for three-dimensional metrology and inspection in the semiconductor industry. The present invention provides such an improved machine vision system.

1 is a bottom perspective view of a system according to a preferred embodiment of the present invention.

2 is a bottom perspective view of an illumination subsystem in accordance with a preferred embodiment of the present invention.

3 is a view schematically showing the configuration of a controller according to a preferred embodiment of the present invention.

4 is a bottom perspective view of a system in accordance with a preferred embodiment of the present invention.

The preferred embodiments according to the present invention presented below are not intended to limit the scope of the present invention to these embodiments, but to enable those skilled in the art to (easily) make and use the present invention.

As shown in FIGS. 1 and 2, a system 10 according to a preferred embodiment includes a structural subsystem 20 that provides a stable platform and vibration isolation; Camera subsystem 30 for capturing an image of a moving object as the object moves to a viewing area, and lighting subsystem enabling obseravtion of reflectance of complex objects under different conditions subsystem 50). The system 10 specifically illuminates a moving object, such as connecting terminals (eg, lead, ball grids, and pads) of packaged electronic components packaged in a manufacturing facility ( is designed to illuminate) and capture its image. The system can, however, be used to illuminate and capture images of any suitable object at rest or moving.

The structural subsystem 20 according to the preferred embodiment functions to provide a stable platform for the relative position between the camera subsystem 30 and the lighting subsystem 50, and furthermore, the camera subsystem ( 30 and illumination subsystem 50 function to isolate from vibration. The structural subsystem 20 preferably comprises an enclosure 22 and a back plate 24. The enclosure 22 functions to provide a relatively rigid connection to the ground or other suitable fixed structure. Preferably, the enclosure 22 is placed or seated in the mechanical structure of a machine that manufactures, inspects, and / or packs semiconductor or packaged electrical components. Alternatively, the enclosure 22 may be placed or seated in any suitable location and suitable device or structure. The back plate 24 functions to provide a relatively firm connection between the camera subsystem 30 and the illumination subsystem 50, and the enclosure 22 and the camera subsystem 30 and the illumination sub System 50 functions to provide a relatively flexible connection between the assemblies. The back plate 24 is preferably made of light weight, high thermal conductivity heavy gauge anodized aluminum. The high thermal conductivity of the back plate 24 promotes heat transfer from the camera subsystem 30, which otherwise guides errors into the system. The back plate 24 is preferably suspended against the enclosure. Such suspension preferably comprises a single connecting plane, but alternatively any other that provides vibration isolation and mounting distortion isolation between the back plate and the enclosure. Mounting methods or instruments. Isolating the camera subsystem 30 from sources of vibration, mounting distortion, and other errors increases the accuracy of the system, thereby reducing the use of the system 10 for three-dimensional measurement and inspection in the semiconductor industry. To facilitate.

The camera subsystem 30 according to a preferred embodiment is mounted on the back plate 24 of the structural subsystem 20 and performs a function of capturing an image of a moving object when the object moves to the viewing area. . The cameras are preferably fixedly mounted so that they vibrate together when they vibrate. Preferably, the camera subsystem 30 includes a first camera 32, a second camera 34, and a third camera 36 to provide information at various angles with respect to the moving object. The camera subsystem 30 may, however, include any number of cameras suitable for providing information at various angles with respect to the moving object. Each camera is preferably a CCD-type camera having a resolution of at least 4 megapixels (MPixels) in 12-bit grayscale and having a field of view of at least 50 mm x 50 mm. Each camera, however, may be any type of image capture device having a suitable resolution and field of view. The cameras are preferably arranged relatively symmetrically with respect to the object in the viewing field, and each camera preferably has a unique viewing angle in the field of view. In a first variation, the first camera 32 and the second camera 34 preferably have complementary acute and obtuse angles (eg 60 ° and 120 °) while the third camera Preferably has an orthogonal viewing angle (ie 90 °). In a second variation, the cameras are spaced radially equally about the Z axis (eg 60 ° for three cameras) while all cameras have the same viewing angle (eg 75 °) about the Z axis.

Based on the use of images captured in a single time instance from three pairs (or more) cameras that are symmetrically arranged, the decision about the three-dimensional measurement of the object is based on statistical error ( based on the reduction of statistic errors (caused by randomness and hence implicitly unpredictable, fluctuations in the measuring device) and systematic errors (caused by unknown or non-random fluctuations) This results in a remarkable improvement in the accuracy and repeatability of the cischem 10. The use of images captured in a single time interval prevents any movement or vibration of the object from significantly reducing the accuracy of the system. The use of symmetrical arrangements eliminates or at least reduces systematic errors (such as one camera moving slightly after correction) because errors resulting from opposing camera pairs are eliminated. The use of three pairs (or more) of cameras tends to eliminate or at least reduce the statistical error since the three-dimensional measurement of the object can be calculated three or more times. For example, the following three camera pairs: C1 and C2; C1 and C3; And C2 and C3. The three decision values obtained from the three sets of cameras are averaged (or statistically analyzed according to other suitable methods), randomly and thus inherently unpredictable fluctuations in the measuring device. Can significantly reduce the statistical error caused by

As shown in FIG. 1, the system 10 according to a preferred embodiment also includes a mirror assembly 40. The mirror assembly 40 performs a function that enables miniaturization of the system 10 and facilitates a specific viewing angle of the first camera 32 and the second camera 34. Preferably, the mirror assembly 40 includes a first mirror 42 optically folding the field of view of the first camera 32 and a second mirror 44 optically folding the field of view of the second camera 34. ). The mirror can be placed between the object and the lens as shown, or the mirror can be placed between the camera and the lens.

As shown in FIGS. 1 and 2, the illumination subsystem 50 according to a preferred embodiment performs the function of irradiating a moving object with strong light having a short wave. The illumination subsystem 50 preferably includes a first illumination group 52 having a light source aimed towards the viewing area formed along the first camera 32 and a viewing angle of the second camera 34. A second illumination group 54 having a light source aimed towards the viewing area formed according to the second illumination group 54 and a third illumination group 56 having a light source aimed toward the viewing area formed along the viewing angle of the third camera 36. Include. Additional lighting, either direct or indirect, may be provided (see FIG. 4). Preferably, the illumination subsystem 50 is seated in a trapezoidal shaped structural member 58, wherein the camera subsystem 30 and the structural member 58 are arranged so that cameras may be mounted on the structural member 58. Are arranged to face the holes formed in The structural member 58 is preferably mounted directly or indirectly on the back plate 24. The lighting subsystem 50 is, however, mountable to any suitable structural member. The lighting subsystem 50 preferably includes at least eight high intensity LEDs for each lighting group, but may alternatively include a suitable number of light sources.

As shown in FIG. 3, the controller 60 according to the first embodiment is connected to the camera subsystem 30 and the illumination subsystem 50, and the camera subsystem 30 and the illumination subsystem ( 50) enables observation of the reflectance of illumination by objects moving under different conditions. In a first change, the controller 60 controls the movement of the illumination subsystem 50. Based on the information collected by the machine vision subsystem (such as the position of the object read from the encoder), the controller 60 can (1) in a fast or slow cycle (eg 50 ms), (2) in the cycle The illumination sub for imaging at (4) and / or high or low intensity (eg 0-50 amps) for long or short durations (eg 5-50 ms) at or early (eg +1 ms) at The movement of the light source of the system 50 can be adjusted. By manipulating the movement of the camera subsystem 30 and the illumination subsystem 50, the system 10 according to the preferred embodiment effectively freezes and inspects the illumination reflectance by a moving object. )can do. The controller 60 may, however, adjust the appropriate factors of the camera subsystem 30 and / or illumination subsystem 50 to enable observation of the reflectance of illumination by moving objects under different conditions. .

The calibration of the system 10 preferably includes a reference target. The reference target is preferably a glass target (available from Max Levy Autograph, Inc. of Philadelphia, Pennsylvania) with a grid of dots. The reference target with a dotted scale serves to provide a position along two axes (such as the X and Y axes). The calibration of the system 10 preferably further comprises a micrometer. The micrometer mounts the reference target and functions to move along a third axis (such as the Z axis). The micrometer preferably comprises a user interface, such as a knob, to enable manual adjustment of the reference target. However, other methods may include automated instruments or processes. The prior art includes a three-dimensional target or a multi-axis robot capable of moving a single dot along the X, Y and Z axes. By eschewing the prior art, the calibration of the system 10 significantly reduces the calibration costs of the system. The calibration of the system may, however, use conventional techniques and / or combine calibration correction techniques to increase accuracy. In observable three-dimensional measurement capacity, determining the three-dimensional shape and position of an object by using a reference target as part of a calibration technique for calculating the correct position for each camera of the camera subsystem 30 according to the coordinate system. To this end, the position of the cameras can be used to triangulate the object directly within this geometrically modified three-dimensional viewable volume.

From the foregoing description, drawings, and claims, it will be apparent to those skilled in the art that various changes and substitutions can be made from preferred embodiments of the invention without departing from the scope of the invention as set forth in the following claims.

Claims (10)

A structural subsystem comprising an enclosure mounted at the bottom of a manufacturing facility, and a back plate extending from the enclosure, the extension of the backplate from the enclosure such that the backplate is isolated from movement of the enclosure; A camera subsystem for capturing an image of the moving object as the object moves into a viewing area, the camera subsystem comprising a first camera mounted to the back plate, a second camera mounted to the back plate, and the back plate A third camera mounted to the; And Receive images captured from the cameras, the first determination value for the three-dimensional measurement of the moving object based on the images captured from the first camera and the second camera, and the first camera and the third camera A second determination value for the three-dimensional measurement of the moving object based on the image captured from the second, and a third for the three-dimensional measurement of the moving object based on the image captured from the second and third cameras A processor for calculating a decision value, The structural subsystem, camera subsystem, and processor cooperate to reduce statistical and systematic errors of the machine vision system to enable accurate three-dimensional measurement and inspection of rapidly moving objects. Machine vision systems for three-dimensional measurement of semiconductor or packaged electronic objects that move rapidly at various stages within a manufacturing facility. The method of claim 1, The processor may further calculate a final decision value for the three-dimensional measurement of the moving object based on the first, second and third determination values for the three-dimensional measurement of the moving object. Machine vision system. The method of claim 2, The processor further calculates a final decision value for the three-dimensional measurement of the moving object based on an average value of the first, second and third determination values for the three-dimensional measurement of the moving object. Machine vision system characterized by the above. The method of claim 1, And each camera of said camera subsystem has a unique viewing angle for said moving object. The method of claim 1, And the cameras of said camera subsystem are arranged symmetrically with respect to said moving object. The method of claim 1, And the back plate provides a relatively rigid connection between each of the cameras. The method of claim 6, And the back plate is made of heavy gauge anodized aluminum. The method of claim 1, And a lighting subsystem for illuminating the moving object. The method of claim 8, The first camera image the moving object at a first viewing angle, the second camera image the moving object at a second viewing angle, and the third camera the moving object at a third viewing angle Is formed to capture the image, The lighting subsystem, A light source in the first light group aimed towards the viewing area along the first viewing angle, A light source in the second light group aimed towards the viewing area along the second viewing angle, and And a light source in the third group of light that is aimed toward the viewing area along the third viewing angle. The method of claim 8, And a controller coupled to the camera subsystem and the illumination subsystem, the controller for controlling the camera subsystem and the illumination subsystem to effectively freeze and inspect the moving object.

Images