Resizing the sky on a multicore platform | Enea Romania | 0LOVES1
>_Loading...

Blog

We're not saying it's gonna be easy.
But we promise it's not gonna be boring.

We pride in providing new challenges all the time for our creative, rebellious engineers. This is the kind of place where we want to change the world with the right lines of code. Insane passion, curiosity and thirst for breaking barriers and rules feel like home here at Enea.

Would you feel like home?

Resizing the sky on a multicore platform

Share

As far as we can imagine, we have gazed at the stars since the dawn of humanity. Hundreds of thousand of questions have been asked since then regarding what’s out there. However, there have only been a few decades since we first reached beyond Earth’s sky and started to gaze at infinity. It was about the same time that we also started to take the first photos.

On the 24th of October, we will be celebrating 70 years since the sub-orbital flight that took the first images from space, a series of beautiful black and white shots that were taken from an altitude of 105 kilometers. 13 years after that moment we had taken the first satellite photos. Countless initiatives have followed after that, amounting to huge data in collected images and this is still happening as you read these lines.

Resizing the sky on a multicore platform
Managing a volume of data and images of such magnitude is one of the biggest challenges of our present times and a team of Enea engineers is working in Iași to develop a solution.

Earth’s sky is rich in satellites, some launched decades ago, with a processing power that is in some cases ten times slower than your average everyday smartphone. This technology lag is caused by the very intensive and time-consuming hardware qualification process needed to make sure a certain processor board will be reliable in space. After all, it’s pretty hard to replace a broken image sensor in space so you want to be sure you don’t need to do that. However, this technology gap means that processing power of the images taken by satellites is very slow, since it’s based on older, single-core platforms, and this hinders the research projects.

This brings us to the fact that just now the first multicore processing boards are certified for space missions, roughly ten years later that when they first appeared, and Enea is the first company to do a research project for the European Space Agency to implement and measure the performance of a special lossless image compression algorithm that will be implemented on a new multicore space hardware setup. Highly skilled in parallel programming and data processing algorithm implementation on a parallel programming framework and AMP operating system, the Enea team working in Iași will also test and measure this algorithm, by making it process images taken by satellites.

Download project case study
First Multicore Implementation for Spacecraft On-Board Image Compression

In simpler terms, it’s all about working faster. When it comes to implementing data processing algorithms – image compression being a very good example – utilizing a multicore platform makes a lot of sense as it can split the processing to all available cores and make it happen in parallel. For example, each core compresses one quarter of each picture or each core compresses a different picture. The overall performance of an X-cores system will never be X times better that a 1 core system, but with advance software engineering, such as optimized RTOS/OSes or parallel processing frameworks, we can get close to X times better.

Resizing the sky on a multicore platform
The gain also comes in terms of volume, since a compressed image using these processing algorithms can be reduced in size up to five times, with lossless compression, which means absolutely no data is lost.

This technology will be fundamental for the next generation of satellites orbiting Earth, used either for space photos or for GPS maps, wildlife tracking and even to law-enforcement.

Related reads

Share

Comments

http://0loves1.enea.com