Machine Vision Applications in Plastics Injection Molding

In the manufacture of consumer products, qualityThese latter quality problems are often not easy to
control is of utmost importance. In most instancesdetect in the processing of plastics. In this paper, the
where manual processes are involved, the individualapplication of a smart vision system in the automatic
who operates the machinery also performs this task. Indetection of poor or improperly oriented print templates
injection molding, for instance, this means ensuring theis presented. The method used to detect improper
correct orientation and quality of print on theorientation is based on edge detection and feature
manufactured parts and components. In manuallyidentification. More advanced algorithms like
operated equipment, this inspection process presents asegmentation, template matching, and character
challenging task because often, after prolonged hoursrecognition are utilized for ensuring proper print quality.
of work, fatigued operators will insert labels inside theTo achieve machine vision successfully, and implement
molding dies in improper orientation, resulting in rejectedthe use of computer software, the viewing area must
parts. In other cases, because of the heat in thebe represented in digital form.The purpose of image
molding process, if there is excess ink on the labelingacquisition therefore is to capture the optical data and
equipment, the prints may be smeared or distorted.change it to a form that will facilitate convenient and
Machine operators occasionally miss or overlookefficient processing using a computer. An image is
quality issues. To alleviate this inspection and operationtypically embedded within a viewing area covered by
problem, it is now possible to incorporate visionthe video capturing mechanism or sensor. The most
systems which will ease the inspection task for thecommon type of capturing mechanism is the charged
operators, allowing them to concentrate more on thecoupled device (CCD). When a light source hits an
manufacturing process. Vision systems provide meansobject, it is reflected to the sensor through appropriate
by which continuous and autonomous inspection canlenses. The photons cause electrical charges to be
be achieved during production.created in the CCl), thus generating analog signals on a
Computer vision techniques are increasingly being used2-D array. The intensity of the charge at each discrete
as alternative methods to conventional inspection andpoint in the 2-D array is proportional to the photon
monitoring applications, due to their simplicity and easeenergy impinging on that point, determining the
of set-up, relative insensitivity to ambient noise,brightness or intensity of the light. Therefore a typical
noninvasive means of gathering information on objectsoptical image system is a continuous 2-D function,
without contact, marking or specimen preparations, andh(x,y) whose value at any pair of spatial coordinates
the potential for online applicability.1"3 Machine visionrepresented by the Cartesian coordinate system (x,y)
has continued to play a major role in the integration ofis the intensity of light at that point. The continuous
automated manufacturing, both from a quality andfunction h(x,y) must be quantized (or digitized) so that it
inspection perspective to more advanced applicationscan be easily processed by computer.
such as motion control and robot guidance. ForThe most common method of digitization is a
example casting mould,mold making,plastic injectioncombination of spatial and amplitude quantization in
mold etc.The role of machine vision in providingwhich the viewing area is divided into a matrix of m by
solutions to manufacturing automation has beenn cells or pixels: m and ç are integers. The image is
recognized by both researchers and engineers insampled at these discrete points in the viewing area.
industry.Each pixel is then assigned a numerical value that is a
In the plastics manufacturing industry, vision applicationsdigital representation of the initial analog value. h(x,y).
have been used mainly to inspect the quality ofthrough the analog-to-digital (A/D) conversion
molded parts, especially for missing features or badlysystem.The numerical value will depend on the bit
formed sections. Although fairly automated, the plasticsresolution of the A/D converter which in turn
molding industry still has a large component of itsdetermines the number of gray-scale levels in which
operations carried out manually, especially for medium-the image can be represented. Therefore the new
and low-volume rates of production. For instance, theimage data, I(m.n). will have values between O (dark or
placement of print template or metallic parts to belowest intensity) and 2^sup r^-l. where r is the bit
included in the finished product is still done manually.resolution of the A/D converter.The acquired image
Under such circumstances, it is possible to have amust then be filtered to remove any noise, and
wrongly inserted print template which would result inenhanced for analysis.This stage is known as
the rejection of the part.processing. There are five common types of
In some cases, due to process variations, a particularoperational approaches for primary processing of
plastic part may posses a smudged or faint print.pixels.