Electronics Information Home arrow Electronics Information arrow Machine vision system
Machine vision system
Thursday, 21 September 2006

Machine vision technology deals with the automated analysis of an image to determine characteristics of objects and other features shown in the image. A machine vision system provides automated, computer-based image processing capabilities that can be customized for various vision tasks, for example, machine guidance, part identification, gauging, alignment, or inspection tasks. Machine vision technologies rely upon image processing and image analysis systems. Image processing systems manipulate raw video images through such processes as filtering, subtraction, and rotation, to generate other, enhanced images as output. Image analysis systems extract information from raw or processed video images to generate information or make decisions regarding the images. Machine vision systems are widely used in many different types of applications, including applications in the fields of industrial automation, process control, test and measurement automation, and medical imaging, etc. Machine vision is often employed in automated manufacturing lines, where images of components are analyzed to determine placement and alignment prior to assembly. The use of advanced machine vision systems and their underlying software is increasingly employed in a variety of manufacturing and quality control processes. Machine vision enables quicker, more accurate and repeatable results to be obtained in the production of both mass-produced and custom products. For example, it can be used in the semiconductor device industry, for example, to analyze images of boards, chips and other components to verify their integrity. It can also be used in the paper industry to detect blemishes on or near the surface of the paper.

Machine vision systems are comprised generally of a lighting system to illuminate a specimen and a camera for capturing light reflected. A digitized image is formed from the light received by the camera. Basic machine vision systems include one or more cameras, which typically having solid-state charge couple device (CCD) image sensors, directed at an area of interest, frame grabber/image processing elements that capture and transmit CCD images, a computer and display for running the machine vision software application and manipulating the captured images, and appropriate illumination on the area of interest. Some vision systems include a programmable illumination system and a lens turret with lenses of various magnifications, for example, in order to increase their versatility and provide the ability to rapidly change their configuration and imaging parameters in order to perform a wide variety of inspection tasks. Optically based machine vision systems, when used for the purpose of mapping the geometry of a surface, typically use one or more of either electronic cameras or input perspectives (using mirrors) in order to create a difference in viewing angle for the scanned subject. Precision machine vision inspection systems can be used to obtain precise dimensional measurements of inspected objects and to inspect various other object characteristics. Such systems may include a computer, a camera and optical system and a precision stage that is movable in multiple directions to allow the camera to scan the features of a workpiece that is being inspected. In machine vision systems, lighting is crucial to the success of machine vision application. In order to increase accuracy and maintain consistency in the data collected by machine vision systems it is important to have a direct source of light to fully illuminate the area being scanned. A large number of lighting or illumination methods exist that are applicable to machine vision. Some of the commonly used lighting methods can be classified into three categories: back lighting, front lighting, and structured and filtered lighting. Vision systems typically require a light source with full visible-spectrum light emission. The light source produces light and the light is reflected off of an object to be inspected.

Machine vision systems acquire images of an environment, process the images to detect objects and features in the images and then analyze the processed images to determine characteristics of objects and other features detected in the images. The system generally includes a camera/frame grabber system that generates an image that consists of a plurality of digitized image pixels. The image pixels are then processed with an algorithm implemented in software and/or hardware typically called a vision "tool." In machine vision applications, a vision processor communicates with a host controller, computer, or other processor to automate such tasks as surface mounting of semiconductor devices. The communication between the two devices typically is in the form of commands the vision system is to perform and responses from the vision system regarding the results. Users of vision processing systems often need to customize or modify the vision processing features and may also develop new features for which new commands must be created. General purpose precision machine vision inspection systems are also generally programmable and operable to provide automated video inspection. The programming capability allows an automatic inspection event sequence to be defined by the user for each particular workpiece configuration. Such systems typically include features and tools that simplify the programming and operation of such systems, such that operation and programming can be performed reliably by "non-expert" operators. The programming capability also typically provides the ability to store and/or output the results of the various inspection operations. Such programming can be implemented either in a deliberate manner. The use of advanced machine vision systems and their underlying software is increasingly employed in a variety of manufacturing and quality control processes. Machine vision enables quicker, more accurate and repeatable results to be obtained in the production of both mass-produced and custom products.

Machine vision applications may use image processing software operable to perform various types of image analysis or image processing functions or algorithms in examining an acquired image of an object. Central to many machine vision applications is the comparison of patterns and images. For example, pattern matching algorithms are often used in order to compare the pattern information of the acquired image to the pattern information of a template image. A common machine vision technique for comparing a pattern and an image is golden template comparison, or GTC. This involves aligning the pattern with the image and subtracting the two. Color matching algorithms may also be used in order to compare the color information of the acquired image to the color information of a template image. Blob (binary large object) analysis tools may also be used to examine an image. These image analysis functions or algorithms may be used to verify that: an object includes all necessary components in the correct locations, an object has the appropriate words, labels, or markings, an object surface is not scratched or otherwise defective, etc. In machine vision, image enhancement techniques are used to process image data to facilitate operator and automated analysis. Commonly known image enhancement techniques can be divided into two broad classes: point transforms and neighborhood operations. Point transform algorithms are ones in which each output pixel is generated as a function of a corresponding input pixel. Neighborhood operations generate each output pixel as a function of several neighboring input pixels. In many machine vision applications, dilation and erosion software tools are used to emphasize or de-emphasize patterns in digital images and to facilitate the recognition of objects in them. The dilation tool is used to enlarge features in an image. Application of this tool typically enlarges and emphasizes foreground surfaces, edges and other bright features. The erosion tool does de-emphasizes bright features by eroding their borders. Machine vision has traditionally been performed on gray-scale images, rather than color images. Gray-scale images acquired of components on an electrical equipment assembly are typically processed using gray-scale machine vision tools, that determine component orientation and location. used to convert a gray scale image. Thresholding is an image enhancement technique for reducing the number of intensity, brightness or contrast levels in an image. Thresholding is commonly used in machine vision systems to facilitate detection of defects. Color video cameras are gradually replacing gray-scale cameras as conventional image acquisition devices. Color video cameras generate digitally-encoded images in which each pixel is rendered in not one, but multiple, spectral bands.

Machine vision systems are frequently used in contexts that require the system to capture a two dimensional image of a physical object and locate within that image some aspects or features that are to be analyzed, such as to determine the position of the features, inspect the integrity of the feature, or effect alignment or a particular orientation of the physical object based on the features found. In many machine vision applications, machine vision analysis is required of a three dimensional or 3D object. For example, wheels of motor vehicles may be aligned on an alignment rack using a computer-aided, three-dimensional (3D) machine vision alignment apparatus and a related alignment method. Cameras of the alignment apparatus view the targets and form images of the targets. Information obtained from each camera is then used to determine the relative positions and orientations of the cameras. Since each camera indicates where the target is with respect to itself, and since each is viewing the same target, the system can calculate where each camera is located and oriented with respect to the other. This is called a relative camera position (RCP) calibration. A computer in the apparatus analyzes the images of the targets to determine wheel position, and guides an operator in properly adjusting the wheels to accomplish precise alignment. To determine the alignment of the motor vehicle wheels, such 3D aligners use cameras that view targets affixed to the wheels. These aligners generally require a calibration process to be performed after the aligner is initially installed at the work site. One important image processing capability typically provided by machine vision systems is object edge detection. Object edge detection is generally achieved by extracting selected information from an object image and evaluating the selected information to locate edge points within the object image.

The advent of increasingly faster and higher-performance computers, has enabled the development of machine vision systems that employ powerful search tools. Searching is a fundamental operation in machine vision. It is used to determine the location of a mark, object or other "template" in an image. Search tools enable a previously trained/stored image pattern to be acquired and registered/identified regardless of its viewed position. Machine vision tools acquire an image of a pattern via a camera and analyze the outline or a particular part of the pattern. The search tool determines the coordinates within an image reference system for each analyzed point in the viewed area, and correlates these through repetition with a desired pattern. The search tool may map the locations of various points in the captured image to stored points in the model image, and determine whether the captured image points fall within an acceptable range of values relative to the model image points. Most search tools can register such patterns transformed by at least three degrees of freedom, including two translational degrees (x and y-axis image plane) and a non-translational degree (rotation and/or scale, for example). The rotation/scale-invariant search (RSIS) tool registers an image transformed by at least four degrees of freedom including the two translational degrees (x and y-axis image plane) and at least two non-translational degrees (z-axis(scale) and rotation within the x-y plane about an axis perpendicular to the plane). Some tools also register more complex transformations such as aspect ratio. Using various decision algorithms, the tool decides whether the viewed pattern, in a particular rotation and distance (scale) corresponds to the desired search pattern.

Machine vision is currently utilized throughout commercial industry in a wide variety of applications. Digital data and signal processing techniques and vision system technology have tremendously advanced the ability to use computers as data processing systems to accomplish sophisticated inspection procedures without human intervention. Machine vision systems are increasingly employed to replace human vision in a wide range of processes such as manufacturing operations. Machine vision systems are often used to inspect objects to determine characteristics, abnormalities or defects in the object. Machine vision is obtaining increasing significance in industry to aid in robotic assembly systems as well as inspection systems for product sorting or quality control. For example, machine vision technology is used in semiconductor wafer fabrication, pharmaceutical manufacturing, food processing, mailpiece processing, circuit board assembly, and many other areas. Machine vision has also been employed in varying degrees to assist in manipulating manufacturing engines in the performance of specific tasks. One task using machine vision is visual servoing of robots in which a robot end effector is guided to a target using a machine vision feedback. In the semiconductor and electronics industries, machine vision is used to determine the position and orientation of semiconductor chips and other components before they are soldered into place on printed circuit boards. Machine vision technology is widely used to inspect the sealing surfaces of glass containers as they are being manufactured or for reuse, to automatically reject defective containers.