The App Store from Apple allowed people from all over the world to share applications with each other, creating a database of easily accessibly tools, games and programs for Apple users. Many have followed in Apple’s footsteps, with application databases for everything from phones and tablets to Internet browsers and computer desktops. It seems that for most things in life it now holds true that “there’s an app for that.” Our phones can scan barcodes, track our exercise habits, suggest dinner recipes and remind us to pick up the groceries. The endless possibilities of what your phone or tablet can assist you with make it an invaluable tool for many people in today’s world.
It’s that type of flexibility and quick access to pre-created tools that Tordivel is trying to re-create with their machine vision app store. The have launched a commercial machine vision app repository to enable OEMs to build dedicated vision applications. You can find the store at http://www.scorpionvisionappsforsony.com, and many of apps are free to download. The first apps? Component alignment, barcode scanning, color identification and object location. So far all the apps are for Sony Smart Camera’s only.
Thor Vollset, CEO of Tordivel said, “To enable as many OEMs as possible to benefit, it was very important that this was created as an open platform. Machine vision is being adopted by a wider range of companies to help in manufacturing or being integrated their products. Many of these solutions only require one or two specific image processing features. With the Scorpion Vision apps model these solutions can be implemented quickly and inexpensively.”
The app repository is also open to third party application developers seeking to expand their user base for both niche and common usage apps. (See original article for more information).
This type of sharing can be a really good thing. However, we do have a few questions that would need to be addressed for this machine vision app store to be successful:
1. Would this only work for low-level applications? In other words, most machine vision engineers spend a lot of time calibrating the camera to the specific product and teaching it to find certain things. Can this level of sophisticated calibration and training really be captured in an easy-to-use app?
2. How could app integrity be validated? Is someone going to make sure these apps really work before they become available in the store? There is nothing more frustrating that thinking you have a solution only to find out that it doesn’t perform very well or that it’s a scam.
3. Where would you go for troubleshooting support? With traditional machine vision applications, the company who provided the machine vision system also provides trouble-shooting support and technical guidance. If you are downloading an anonymous app, would there be any support?
4. Will it become available for other cameras? For this to really be successful it would probably need to include apps for multi-camera brands.