Combining open source tools with in-house expertise can open up a world of possibilities for highly tailored solutions.
The ‘not invented here’ bias is a fascinating psychological principle where in-house solutions are preferred over third party products. Ignoring the negative connotations that are usually associated with this bias, there are a few reasons why a company might prefer to solve a problem themselves rather than use a piece of software, or equipment, created elsewhere. The external solution might not meet all their needs; it might not be adaptable to changing requirements; be secure enough; or it might introduce a reliance on proprietary IP. Another genuine concern is the potential high purchase price.
However, external solutions have many benefits that often outweigh these concerns. By importing a ‘turnkey’ external solution a company can take advantage of external knowledge and experience that they don’t currently possess. It is true that external solutions can be expensive. But it is all too easy to neglect the time and effort required to turn a simple solution into a reliable and durable system that is genuinely easy to use by all, and not just to the person who created it! Not to mention the cost of installation, maintenance and enhancement - this is all covered externally.
Crucially, not all external solutions cost money, there is a growing amount of open source solutions collaboratively created, maintained and improved by groups of really clever and creative people from around the world. What’s more many are completely free!
At Cambridge Consultants we are technology agnostic - we are not afraid to consider solutions generated elsewhere, and incorporate them into our products. By combining open source external tools with our diverse in-house expertise we can develop advanced solutions more rapidly and cheaply than if we were to develop everything ourselves. What’s more, we can bring together tools from many different disciplines to develop a solution tailored specifically to the needs of our clients.
Working as a Physicist in our Sensing Systems group, I spend a lot of time developing algorithms to extract useful information from complex physical systems. Recently I and others within the group have been concentrating on combining image processing with Cambridge Consultants’ existing expertise in robotics and controlled dispensing. Our work builds on some fantastic external open source tools such as OpenCV, Caffe and scikit-learn, to name but a few. We are using neural networks and machine learning to identify objects from video streams, in real time, and physically act on that information using Bayesian decision making. We have two concept demonstrations which we will be exhibiting for the first time this year at Agritechnica 2015. We look forward to hearing how we can help build these concepts into your products to solve key industry problems.