Case study / Industrial imaging and ERP

Image analysis and project management modernisation for an industrial manufacturer

Engagement: 2005-2007Sector: Textiles and manufacturingStack: VB.NET, ASP.NET, SQL Server

A textile manufacturer in southern Africa needed two pieces of software the off-the-shelf market did not provide: an image analysis system to evaluate carpet samples programmatically, and a modernisation of their internal project management platform from legacy classic ASP into ASP.NET. We delivered both as part of an early-career engagement that taught the patterns we still apply to manufacturing software today.


The brief

The client was a major textile manufacturer with two distinct technology problems. The first was the absence of a way to analyse carpet sample images programmatically — an internally-driven need around quality and pattern evaluation that no off-the-shelf product addressed for their workflow. The second was a legacy project management system written in classic ASP that had grown beyond what the original codebase could maintain, and needed lifting to a modern ASP.NET stack without breaking the work in flight.

Image analysis system

Built in VB.NET against the .NET Framework available at the time, the image analysis system processed carpet sample images for the manufacturer's internal evaluation workflow. Pixel-level analysis, pattern recognition routines, and reporting outputs that the production team could action on the floor rather than wait for back-office review. A useful early lesson in the gap between research-grade image analysis and production-line reliability.

Project management system conversion

The legacy ASP project management system, plus its SQL backend, lifted to ASP.NET with a modernised data layer. The conversion preserved the existing business logic the team relied on while moving the front end to a maintainable modern stack. Done as a working modernisation, not a rebuild — the team's day-to-day operations continued on the modernised system without retraining.

What we learned, applied across our other work

Custom imaging beats off-the-shelf when the workflow is custom. Building against the manufacturer's specific evaluation criteria taught the pattern we apply now to document AI and OCR work: the libraries are there, the production discipline is not.

Modernise in place; rebuild only when the math is right. A working classic ASP system that the team relies on every day is a higher-value asset than a hypothetical greenfield rebuild. This is now our default position when clients ask "should we rebuild it." Usually not.

Industrial software has different reliability constraints than office software. A bug in a back-office reporting tool is annoying. A bug in a manufacturing-floor system can stop a production line.

Tech stack

Image analysis

VB.NET.NET FrameworkCustom imaging routines

Web modernisation

Classic ASP → ASP.NETHTMLJavaScript

Data

SQL ServerT-SQLADO.NET

Want this kind of outcome?

We work with mid-market companies across the UK, US, and Australia. Voice AI, Finance AI, Data AI — and bespoke for everything else.