Rewarding sides of a project other than only the technique, what pays off?
Being a technical consultant for the last 25 years, it is sometimes also good to look back on a project with a broader view and see all the good things happening besides the tech-stuff. So, I leave most of the technique for another post and will mostly focus on all the other things that I have experienced during the VidaXL project I have been working on lately.
A clear explanation
First, I must explain you what we have done exactly in this project for our client VidaXL. VidaXL is an online only shop that sells everything for your home, garden and even more. Their slogan: ‘Live it up for less’.
They have huge warehouses where hundreds of thousands of products are stored. It is a big challenge to keep track of all these goods. For example, when boxes get mislabeled or complete pallets are placed in the wrong location. This is where we, as Heroes, come in. Together we defined two initial use cases namely for goods receiving and the so-called inventory viewer. Because the inventory viewer was the most difficult one, we decided to focus on that one first.
The inventory viewer is a moving installation with several industrial cameras attached to it that will ride alongside pallet racks to detect pallets and locations. On both a detected location-label and pallet it will find and decode the barcodes and texts. In the prototype phase it will only do so on a single shelve level. But how to start?
Standardized approach
At Heroes, we have developed a standard approach for such innovative projects like the one we are currently working on with VidaXL. Our method is divided in five phases. In this project we are now at the end of phase 3, the so-called prototype phase and about to start the MVP phase and this in only five weeks!
The advantage of working with this standardized approach is that at the end of every phase you can come to the conclusion that it won’t be possible to turn the idea into production. This way you know very fast how feasible the use case is and minimize the cost of investment in case it is not.
Teamwork & co-creation
Another great advantage of this method is the boost it gives you as a team. As you are starting from scratch together, you will immediately be on the same page striving for the same results. Actually, it is much more like a co-creation than a traditional client-supplier relationship. The involvement from the VidaXL team provided us with much-needed domain knowledge.
Open for innovation
Although VidaXL knows that there are solutions for this type of use case which you can buy ‘of the shelves’ they decided to take the opportunity and build an, in the end much more flexible and tailored solution, together with us. From all things that pay off most in such projects this is my absolute number one, the will to innovate.
Partnerships
The most challenging part in this project was the camera hardware. It turned out that we had very high demands on what is still possible and affordable. To be more concrete we needed to read a 0.56 mm (barcode) line from a distance of 1,5 meter and a field of view of 2,5 meters. Because we did not have all the knowledge in house, we decided to reach out to getCameras (https://www.vision-camera.nl/) and they could give us a first-time-right advice within hours. We even decided to partner with them for this and future projects. For me it is a reconfirmation that working together is the only way you can ever succeed.
Open-source software as a companion
It is common to use open-source software for computer vision and Artificial Intelligence (AI) and some amazing building blocks are out there. So, we selected a SOTA (state of the art) YOLO model as the base for our pallet and location detection.
Also our Optical character recognition, barcode detection and integration layers are all build on and with open-source software. With the additional benefit that there are absolutely zero license costs.
Kickstart the next phase
We are about to start the MVP phase and I will definitely keep you posted on that. We are very much looking forward using new techniques like the part of AI that also enables the creation of Deepfakes. We will use it only for a good purpose by enhancing the photos we take, so that we need less or cheaper cameras and/or we can build an even more reliable solution.
In the end innovation is our DNA!
