VCVS2

Eltvor's CubeSat-compatible OBC/DPU originating from our successful mission of dual camera. The board contains Zynq 7k CPU+FPGA SoC, connectivity to platform bus is via CAN and RS485. (I²C is populated as well, but not recommended.) Other connectivity includes: 1000BaseTx Ethernet, USB host port. Power consumption is around 1W. The DPU includes 4GB of eMMC non-volatile storage.

The DPU may directly connect 1 or 2 camera modules of our design, which was the original mission. In case of interest, Eltvor has these cameras ready, B&W and RGB (Bayer) versions, 1.3Mpix, 13eRMS readout noise, 10000e well capacity.

The DPU as well as the camera modules do have a flight heritage (photographing right now at 520km altitude), vibration & shock to Falcon-9 + ISIPOD profile, TID >=25krad (3 samples), T-vac.

Please see attached a brief specs sheet of the VCVS2 DPU.

The equipment offered:

The development environment for VCVS2 is standard ARM Linux. If camera option is selected, FPGA firmware is delivered including a C library and a standalone Linux executable which is able to shoot pictures in 1.3Mpix * 10b quality. In flight, we have successfully performed SW updates via CAN, JPEG2000 compression, automated image scoring and exposure control etc.

Figure 1. VCVS2 module

Figure 2. VCVS2 DPUs and VCAM1M3 cameras installed in VZLUSAT2 (VCVS2-EGSE is also displayed)

Figure 3. Photos taken using VCVS2 with VCAM1M3-BW (B&W camera) on-board VZLUSAT2

Figure 4. Photo taken using VCVS2 with VCAM1M3-RGB (RGB camera) on-board VZLUSAT2

On-board computer for Gamma camera

Eltvor's VCVS2 participates on another independently operated instrument on board of VZLUSAT-2: Gamma camera with matrix detector. The physics package as well as application SW was developed by companies Rigaku and AdvaCam, while the on-board computer and OS infrastructure was provided by Eltvor.

Figure 5. Eltvor's VCVS2 used as OBC for Rigaku and AdvaCam Gamma detector

In-flight upgrade of Earth observation camera with neural classifier

A group of young entrepreneurs of Zaitra startup contacted Eltvor with an interest to leverage TRL for their neural network based classifier of Earth images. Their original intent was to get Eltvor's Vitacam computer on a loan and to run some dummy tests. Instead, Eltvor proposed and later managed to upgrade its Vitacam in-orbit, and to run the neural network on Eltvor's computer in flight.

The SW upgrade went well, and the neural classifier was run alongside with Eltvor's original automated image quality classifier. The plot below shows the score calculated by Zaitra's neural network and by Eltvor's original SW, and as an example one of the nice photos taken during this event, depicting an ice field between Greenland and Canada.

Figure 6. Comparion of Eltvor's j2k-based classifier with Zaitra neural classifier

Figure 7. A high-scored picture found by the on-board software classifiers