The inspection process
Four steps of a simple inspection application
A simple inspection application normally consists of the following steps:
Acquiring the image
The goal is to set the system up such that it will acquire the best possible image that highlights the defect. This may require a combination of specialist lights and filters. It must be remembered that a camera is not as flexible as the human eye so ambient lighting is often unacceptable. Adjustments to the system such as setting the time of the camera trigger, adjusting the camera exposure to reduce motion blur, or taking a decision on whether lighting control will be required, might be necessary.
Locating the product
Once the image is captured, the item under inspection needs to be located within the image. It is desirable that the object is in the same location and orientation to the camera for each inspection as this can simplify the inspection. Using a mechanical system to do this is ideal, however, in many applications this is not possible or the mechanical system does not remove all movement. To deal with this the first step is to locate the part in the image. This is often termed creating a fixture or anchor point and usually involves finding two edges or two patterns on the item, such that all further inspection tools can be referenced to these features. Depending on the size of the expected range more time may be needed to accurately locate the object.
Once the image is located the inspection tools are applied. Common inspection tools are discussed later in this chapter, however the complexity of the inspection phase can vary significantly which in turn can cause the inspection time to change.
Decision and communication
The decision is made after all the results from the deployed tools have been evaluated. Once the result is known the physical pass/fail line can be set to the relevant level, or in some applications this reject output needs to be synchronised to operate as the product passes further down the production line. Most modern vision systems now support the same Ethernet control standards as PLCs such as ethernet IP so results and status can be directly linked to PLC Tags enabling tight integration of the vision system and the PLC.