Emergent to Showcase High-Speed GigE Vision Camera Innovations at Vision Stuttgart 2024

eye4you vision stuttgart 2
Vision Stuttgart Ticket Code

Emergent Vision Technologies will showcase its award-winning, high-speed imaging technologies at Vision Stuttgart 2024 from October 8 – 10. In Hall 8, Booth 8D60 of the Messe Stuttgart, Emergent will demonstrate its FlexProc & FlexTrans technology, real-time 3D reconstruction capabilities, its GigE Vision cameras, software, and network interface cards, and much more.

Don’t miss this opportunity to meet the Emergent team in person and learn about how we can address your most challenging AI and machine learning applications. Get your free ticket here and simply use the code EMERGENTVISION for registration. We look forward to seeing you soon!

Below is a look at the different products and technologies attendees can learn about while visiting the Emergent Vision Technologies booth at Vision Stuttgart 2024.

Vision Stuttgart 2024 Demos

eCapture Pro – Real-Time 3D Reconstruction Software Demo

 

eCapture Pro is a drag and drop application where one can add devices to the system through a graphical interface. We add the cameras, servers, storage, GPUs, etc, and then connect them to define the hardware for the system or use the auto connect functionality. We also configure the processing nodes like the GPU to define their behavior which in this case is to do real time volumetric video. For a volumetric video appliciation, a project is created and single multi-camera captures are taken with simple calibration targets present. With this done, we simply click Run to run the application which generates the volumetric video. We can view this in a few ways. We can use a joystick and fly around the volume and view the subject in our 3D space on a 2D monitor. We can also use Meta Quest 3 for example to fly around the 3D space using the Quest controllers and, with the Quest goggles on, can look all around by moving our head. The Quest integration is enabled with the use of Unity. The 3D content could be transmitted to another location for remote viewing.

hr series logo
ecapture pro
hr c mount side to side 1170x600

And all of this in the video is done on a single mid-range server with only 1 GPU. eCapture Pro is a UI based software allowing customers to fully integrate Emergent award winning cameras which have speeds of 5GigE, 10GigE, 25GigE, and 100GigE. The software is often used for large scale recording systems involving multiple servers, switches, storage, etc. Up until now, the software has been limited to recording camera data to storage in a very high performance manner for the purposes of post-3D reconstruction for Volumetric Video applications. Emergent has developed and added to its award winning eCapture Pro software the ability to generate real time Volumetric Video so one can be immediately immersed in the event or subject being filmed. Imagine being in another city with VR goggles such as the Meta Quest 3 and being able to navigate anywhere in the 3D space we provide. Imagine flying around the space using the Quest 3 controllers which control speed and direction. Imagine, as flying around, one simply turns ones head to look in any direction. Or, by all means, find a comfortable view and enjoy the show from that unique perspective.

100GigE HZ-65000 Dual Fisheye Demo

This demo utilizes the 100GigE HZ-65000 camera and the new Canon RF dual fisheye mount. Used along side a VR headset such as the Oculus Quest 2, etc., users can immerse themselves into a vivid, high resolution 3D experience. This lens is designed specifically to deliver outstanding stereoscopic 3D imagery. Featuring dual 5.2mm fisheye lenses positioned with 60mm of spacing between Left and Right Lens centers, it provides an interpupillary distance closely resembling human vision with a natural parallax for a powerful, immersive experience that brings the viewer to feel they are within the scene where the video or stills image takes place. With an impressive 190 degree optical field of view, this provides the viewer the ability to change their viewing angle within the recorded scene, allowing them to look around with use of a compatible head mount display.

canon rf dual fisheye
zenith logo
zenith logo
ecapture pro
ecapture pro
gpu direct logo
gpu direct logo
zero copy logo
zero copy logo
zenith mixed 1170x600 2(1)
Little Planet
zenith mixed 1170x600 2(1)
Little Planet

100GigE HZ65000 Little Planet/Entaniya Fisheye Lens Demo

entaniya lens
entaniya lens
Using Entaniya fisheye lenses, the Emergent EF mount adapter, and remapping algorithms, we are able to show 360 degree panoramic views in a variety of ways. The images are transferred from the camera to the NVidia GPU using GPU Direct and the algorithms are deployed using Emergent eCapture Pro GPU plug-in feature for high performance rapid prototyping. This Entaniya lens is capable of generating a full 360 panorama using only a single image sensor and camera.
zenith logo
zenith logo
ecapture pro
ecapture pro
gpu direct logo
gpu direct logo
little planet
zenith mixed 1170x600 2(1)

Eros/SWIR HE25000-SBL & HE5300-SI Machine Learning Conveyor Demo

detection samples

This demo features two 5GigE HE25000-SBL cameras and two 5GigE SWIR HE5300-SI cameras. Using machine learning libraries such as PyTorch, these models are trained to detect and report defects on subjects such as blister packs. Using GPU Direct, images transfer directly from the cameras to the GPU. The GPU runs the in-house trained machine learning model to results in real time. This demo deploys the algorithms using eCapture Pro plug-in feature for rapid high performance prototyping using zero copy and GPU direct.

Nvidia IGX Orin
Nvidia AGX Orin

This demo can be demonstrated with Nvidia Orin IGX, Nvidia Orin AGX, or on lower end computers based on an Asus motherboard with Intel i7.

eros logo
eros logo
ecapture pro
ecapture pro
gpu direct logo
gpu direct logo
poe logo
poe logo
he rj45
he optical transceiver

100GigE LZ-16KG5 & LZT-9KG5 Line-Scan Demo

This demo has a linescan running with a single rotary stage using its built-in encoder. The LZ-16KG5-M has a 400KHz line rate and a 16Kx16 resolution while the LZT-9KG5 has a 608KHz line rate and a 9K 256TDI resolution. These linescan cameras are the fastest monochrome line scan cameras with a 100GigE interface, scanning a printed sheet of paper on a drum rotating on the vertical axis. Emergent eCapture Pro and eSDK are used for best in class zero copy driver performance. GigEVision 3.0 RDMA/RoCEv2 is also used in this demo seen as the fasted deployed solution for RoCEv2 for the new addition to the GigEVision standard slated for release in 1-2 years.

pinnacle logo
ecapture pro
zero copy logo
rdma rocev2
tlz 9kg5 no fan (4k)

Emergent 100GigE Line-Scan Camera: Pinnacle LZT-9KG5

lz 16kg5 (no fan)

Emergent 100GigE Line-Scan Camera: Pinnacle LZ-16KG5

RoCE Graphic

Comparison of sockets-based transport with RoCEv2 zero-copy RDMA

eCapture Pro – FlexProc & FlexTrans Demo

FlexProc
– User Defined Processing
– Custom CPU, GPU, FPGA Plug-ins
– eCapture Pro and eSDK support
– Endless processing flexibility

FlexTrans
– Any node to any node Transfers
– Zero-Copy and GPU DIrect support
– eCapture Pro and eSDK support
– Endless Scalability

This demo utilizes 8 of the 25GigE HB-25000-SB-C cameras.

Bolt 25 GigE
ecapture pro
gpu direct logo
zero copy logo
zero copy logo
bolt c mount group mix 1170x600

FlexProc allows one to develop and easily distribute processing tasks anywhere in the system. The blocks within the graph view of eCapture Pro can be configured with built-in processing plug-ins or the user can create their own custom plug-ins. For a custom GPU plug-in, one would develop CUDA code to define the task. The code is then compiled into a .dll and this .dll defines the custom block within the graph view of eCapture Pro. With the cameras and blocks all connected within eCapture Pro, the user would simply click run to run the application.

FlexTrans is more than just the ability to move content from one processing node to any other node in the system. It is also about doing this with the highest performance and lowest latency possible. This means using GPU Direct and Zero Copy methods developed by Emergent. GPU Direct is a method to does direct copies and transfers between GPU memory and the network. We can seamlessly transfer high performance data between a GPU in one server to a GPU in another server. Peer-to-peer is a method to do direct copies and transfers between GPUs within the same server. We also can use the same technologies to take data from server storage devices and system memory to other locations within the system. eCapture Pro FlexTrans allows all of this and with no coding whatsoever. In eCapture Pro, one would simply connect the blocks in the graph view to direct where you want the data to go next in the processing pipeline. The performance aspect is quite important here as a system that does not utilize these performance methods can take up to 3x the resources which translates into 3x the cost. Latency in such systems is not tolerable by the application.

This show runs until Thursday, October 10, 2024. We are located at Hall 8, Booth 8D60, Messe Stuttgart. Come drop by the booth to find more about how Emergent can address your most challenging AI and machine learning applications!

Join us at Vision Stuttgart 2024 and experience firsthand the power and potential of our technologies. At Emergent, we’re capturing the future, today.

Performance and reliability proven!

10+ years shipping 10GigE
6+ years shipping 25GigE
3+ years shipping 100GigE

 RDMA ready

Avoid the imitators
Engage the innovators

October 8 – October 10 | Messe Stuttgart – Hall 8 Booth 8D60 | Stuttgart, Germany

Vision Stuttgart 2024