SenSen

Challenges

Because SenSen’s smart cities project involves parking operations, number plate recognition, and driving distractions, Jetson AGX Xavier was required. NVidia Jetson AGX Xavier was specifically needed to acquire more RAM and processing power as needed for this project.

These systems needed to be equipped with 32 GB of Ram, various sized hard drives, and Ethernet ports. This is because it was vital for the six to eight cameras being used on SenSen’s systems to have a high threshold. In addition, 8GB Ethernet was required for a high frame rate and to maintain a high level of network threshold.

Lastly, NVidia GPUs were required for high performance, fast frame rate, and quick video processing. Quick video processing was important for SenSen as it meant that number plate recognition could be achieved on the side of the road.

To meet our client’s requirements, we had to analyse their needs in detail. This would enable us to fully understand what was needed and therefore provide the best possible solution specific to meet SenSen’s business goals.

How We Helped

DiGiCOR provided SenSen with 15 Jetson AGX Xaviers. It was important that DiGiCOR was able to provide this IoT solution to SenSen for a number of reasons. Firstly, SenSen aimed to use these machines in the building and writing of the software that would be used in the smart city project. One of the focuses of the smart city project was saving power. This involved switching off street lights when the sun is out by creating street light sensors. Secondly, SenSen wanted to target smart traffic management. This involved working for government services to detect people who are using their phones while driving and to issue fines. Lastly, SenSen needed these systems for National Heavy Vehicle Regulator. This looks at providing speeds and number plate recognition of vehicles. However, in the situation, SenSen wanted to use mainly focus on using these systems for number plate recognition.

DiGiCOR also embedded each Jetson AGX Xavier with NVidia GPUs. NVidia GPUs were critical for SenSen. First, they needed a high-performing GPU to handle fast frame rates. For example, some of the cameras being used run 24 frames per second. In addition, NVidia GPUs was needed to enable video processing through SenSen’s servers in real-time to make number plate recognition faster.

In addition, each Jetson AGX Xaviers Controller was equipped with 32 GB of Ram and GB Ethernet. This provided the systems with a high threshold which was needed to handle the fast frame rate of SenSen’s cameras.

Results

The Jetson AGX Xaviers equipped with 32 GB of Ram and 8GB Ethernet enabled SenSen to create successful solutions for their smart city project. In this situation, this predominately involved number plate recognition. This is because these systems enabled SenSen to maintain high level of network threshold which was necessary for the cameras being used. Overall Jetson AGX Xaviers enabled SenSen to acquire more RAM and processing power which was needed for the smart city project given the intense workload present.

The NVidia GPUs DiGiCOR provided to SenSen was able to handle the fast 24 seconds frame rate on their cameras. As a result, they were able to process videos in real-time and which allowed number plate recognition to be done on the side of the road. NVidia GPU delivered high performance which meant they were able to process data that they got in a quick and easy manner. In addition, with NVidia GPUs, the cameras capturing and reading number plates happened in seconds rather than minutes.

SenSen expressed that they have worked with DiGiCOR for 10 years and that it has always been a good experience. SenSen also added that the support has been brilliant and that they’re very happy with what we’ve got.