Pickit 3.2: Unlock new levels of intelligence in the 3D Vision Platform
So, what's new?
The Pickit3D Vision Platform keeps extending its capabilities!
New application - Depalletization thanks to DeepAL!
An efficient depalletizing 3D vision system is within your reach now. The new detection engine, DeepAL, adapted to the depalletization of unmixed products, such as boxes, buckets, or crates, allows you to detect parts arranged in any pattern. The DeepAL engine combines both depth and color information into a state of the art of deep learning algorithm. The combination of 2D/3D images allows for more accurate pick-point calculations. Next to that, teaching parts remains effortless, and the detection of your products remains reliable even during challenging scenarios, where parts are being moved or tilted and exposed to changing light conditions. This detection engine is already working successfully at our pilot customers. From now on, you can also benefit from reliable and easy-to-use software for your depalletizing applications.
Enforce vertical picks
To ensure even better support for your depalletizing projects and other applications with a 4-axis robot, we introduce an additional feature for parts picking: enforcing vertical picks. The latest software update takes into account the limitations of the end-effector which cannot vertically tilt. At the same time, it considers the possibility that robot tools may tolerate some tilt without compromising the pick success. As a result, Pickit software provides optimized feedback on the pickable parts, ensuring a high pick-success rate. Thanks to this improvement, you can count on the robust and improved performance of various applications using a 4-axis robot.
Teach box model
This model type is recommended for applications with randomly oriented box-shaped objects of known dimensions, such as billets with a square cross-section. Whilst we could already do so with cylindrical parts and the existing detection engine: Teach, we now added the 'teach box model' based on parts dimensions. All you have to do is to let the system know the size of your boxes, and voila! You do not have to upload CAD files or teach your part by presenting it to the camera. Thanks to that, you can ensure a simple, fast, and reliable teaching of the 3D model of different products relevant to your application.
Flexible pick position
We understand that you need good-quality feedback to optimize the performance of your application. We now added a comment explaining why a given part is pickable in a certain way, just like you could already see feedback explaining why certain parts are unpickable. The combination of both feedback gives you an in-depth insight into how the system works. Thanks to that, you can save time on the possible fine-tuning of your settings.
Introducing an upgraded camera!
M-HD2 – Lighter, handier, and even more accurate
After extensive testing to ensure optimal compatibility, we have added the M-HD2 model to the Pickit family. This camera is smaller and lighter and assures even better accuracy and a wider view angle compared to the previous M-HD model. It is a perfect fit for robot-mounted applications due to its lightweight and compact form. It uses ethernet communication with the Pickit processor. Thanks to added hardware, you can benefit from the Pickit3D Platform in an even greater range of applications!
Get the update
If you have an older Pickit version and would like to try 3.2, check out how you can upgrade your system.
Read 3.2 release notes in full detail or ask us anything on Pickit's 3.2 release by filling out the form below.