Under development for five years and commercially tested, this third-generation version is a VGA sensor with 15 μm pixels in a 3/4" optical format. It is aimed at developers of cameras in industrial automation and IoT systems such as robots, inspection equipment, monitoring and surveillance devices. It leverages Prophesee's neuromorphic vision technology to offer efficient capabilities for a variety of use models, including ultra high-speed part counting, vibration measurement and monitoring or kinematic monitoring for predictive maintenance.
"This is a major milestone for Prophesee and underscores the progress in commercializing our pioneering Event-Based Vision sensing technology. After several years of testing and prototyping, we can now offer product developers an off-the-shelf means to take advantage of the benefits of our machine vision inventions that move the industry out of the traditional frame-based paradigm for image capture," said Luca Verre, co-founder and CEO of Prophesee.
In Prophesee Metavision sensor, each pixel is independent and asynchronous, only activating if it senses a change in the scene, a movement - an event. This proprietary Event-Based approach allows for major reductions of power, latency and data processing requirements imposed by traditional frame-based systems. It enables sensors to achieve much higher DR than commonly associated with high-speed vision. And it allows cost-efficient sensors and systems to record events that would otherwise require conventional cameras to run at 10,000 fps.
The chip is packaged in a 13x15 mm mini PBGA package and is manufactured in a 0.18um CIS process. The packaging was realized by Kingpak.
Still no clear killer driving application on the radar for this yet interesting technology...
ReplyDeleteMaybe we should ask to these guys https://www.imago-technologies.com/visioncam-odm-intelligent-camera
DeleteWe had this same issue with 3D depth image sensors for many years until massive adoption happened. We do not belong any more to the time of imaging for the purpose of photography and video. We are at the very beginning of a transition of imaging for the purpose of sensing. 2D Biometry was a first killer application for sensing imagers then came the 3D depth image sensors SL and ToF. As analyst of the CIS industry I believe in the developpment of this new technology because it does serve this sensing trend. Samsung, did believe in it, and so did many startups like Prophesee. For sure the marketing picture is still very blurry, and we try to play our part to envision the possible killer applications. To make it short this type of image sensor technology is ideal for motion understanding in a machine vision context at ultra-low power. Therefore we believe two makets could open in the short term, one is the industrial markets addressed by companies like Prophesee and Insightness, and then there are consumer applications which are more addressed by Inivation and Celepixel. The end game everyone is also looking at is Autonomous Driving, Samsung is very pushy in this direction, from our analysis we believe this one might be a medium term opportunity. The imbrication of this new image sensor technology with neuromorphic computing is also extremely exciting. The new battleground for the CIS industry is about merging with AI. The event-based approach is natively neuromorphic (asynchronous, spike based) and therefore it has huge potential benefits which are not yet available due to the pace of AI technology developments. Let's give it 2 to 3 years, but once the killer application becomes obvious it will be too late to invest. Pierre Cambou – Yole Développement
ReplyDelete