Another big project


Another success story in which the use of the B4J tool was crucial. The project is a little older, but I think it is interesting to many people.
I was in charge of preparing signature headers for imprinting data on pharmaceutical products. It is a mechanical print, such as on various medicines, where the batch control number and the date of validity and occasionally also the date of manufacture are printed.
In production, many batches of different products were made daily, so these bearing heads of these signature letters were prepared ahead of schedule. because there was so much of it, there were mistakes.
he forms of these elements were plates and round heads. The idea was that it would be possible to check the correctness of the prepared heads with machine vision.
Since I am creative by nature and this was a big challenge for me, I naturally tackled this project with a lot of energy.
The project requirements were:
- the system must be connectable with planning (an existing software solution, by the way also my work - but it was not done in B4J :) )
- connection to the ERP system with data (batch and dates) and data samples (products had different ways of using the control number and dates depending on the customer or regulatory requirements)
- automatic procedures (saved procedures for quickly repeating the control with different input data).
- the system must comply with the CFR21 regulation
The system connected to several different databases - Oracle, MSSQL and local MYSQL.

The hardware used was the Omron XPectia system and purpose-built lighting that was controlled via RS232 communication. Lighting was key to successful reading and a great deal of time was spent finding the right lighting shape and lighting color.
I would not particularly present the Omron XPectia system itself, because there is a lot to write about. The system has its own control computer to which several cameras can be connected, in this case there was only one.
The system communicated with my program via the UDP protocol. The system uses scenes or work procedures that need to be prepared on the system itself when learning the camera. these procedures can then be called and started remotely. Which data is forwarded at the end of the control is defined in the procedure or scene. The system itself is extremely powerful and flexible.
Using the B4J tool, I created an application that connected all of this together. In the application, it was possible to create your own procedure according to the needs of the control. Of course, it all depended on the level of the user.
Briefly, the procedure was as follows:
- read work order plan for each production line
- read data from ERP for the work order in question
- choosing a procedure on my system, based on a sample of data from ERP
- starting the control procedure
- review of returned data from the Omron Xpectia system
- printout of the control label and signature. the label accompanied the prepared elements for the impression, which were on a special carrier
It took about 1.5 years to develop the entire system, about 6 months for the application itself (including other obligations, of course). Of course, it is also a big financial investment.
The system was realized and tested, but unfortunately it was never validated. There are several reasons for this, including the fact that I left this company and went into the development of the device that I have already presented.
As it concerns data and a project of a sensitive nature, unfortunately I cannot provide the video that was recorded for the presentation. I didn't ask for permission, maybe someday I will.

The purpose of this paper is to show how reliable and how powerful the B4J tool is.

Thank you to everyone, the author of this tool, who helped in the realization of this project and sometimes answered some stupid question.

DSC_4288 .JPG
Last edited: