Once an image has been processed the data must be used for some purpose. Data handling options can vary greatly in type and scale but broadly they fall into the following basic categories:
Local IO
Once the inspection routine has run and a determination has been made, the simplest process which follows is simply to "reject" parts which are deemed to be defective or sub standard. This is usually done through simple IO instructions to a PLC or similar control device. In this case there is no data collection required although it is possible to save images of defect parts if required for QC evaluation and analysis. If this is required, then suitable data storage must be provided. Aggregated statistical data is also usually available.
Local storage
In systems where products are graded, there maybe a requirement to store images and data for review. This may be especially true for systems where payment is based on inspection results. In these cases, local data storage may be provided.
During the project development phase it is advisable to discuss storage issues and security that may be required including the need for backup. Given that industrial vision systems generate quite large amounts of data in the form of images, the architecture and storage policies must be completely thought through.
Cloud storage
The latest trend in this area involves uploading the data to a cloud based server. This can be done in real time or as a batch system depending on user requirements. The usual data security issues apply but can be overcome in most instances. The speed of data transfer is of key importance as is the plan for what happens if the data link is lost.
High speed grain inspection. The image of each seed must be processed and stored for possible future retrieval. This forms part of the client traceability requirements.
Every project and application has its own problems and specifications. To find out how Indyn can assist with your special problem, just click here.
