Sign Up
ANPR Software, Product

ALPR on Nvidia Jetson Devices

Plate Recognizer, the leader in Automatic License Plate Recognition (APPR) software, is proud to announce that its license plate recognition On-Premise SDK technology now supports Nvidia Jetson devices, such as the Jetson Nano, TX1, TX2, Xavier NX and AGX Xavier.

What are Nvidia Jetson Devices?

Nvidia Jetson devices are embedded computing boards made by Nvidia. They are computing platforms that are designed for a specific software-controlled task mostly to accelerate machine learning applications like deep learning.  Jetson devices are generally designed to be power efficient.  The boards typically have a processor, a memory unit, an input/output peripheral device that has a dedicated function.

There are many variants of Jetson devices ranging from Jetson Nano, Jetson TX1, Jetson TX2 to Jetson Xavier.  Each Jetson devices has its unique functionality and improved capacity for different AI applications.

  • Jetson Nano is a small and powerful computer for embedded applications that delivers the power of modern AI.  It was designed mainly for hobbyist robotics due to its low price. The final specification show that the board is a sort of power-optimized, stripped down version of a full Tegra X1 system.
  • Jetson TX1 is the first set of supercomputers designed on a module. It has the ability to produce the performance and power efficiency needed for the visual computing applications. It is built around the revolutionary NVIDIA Maxwell™ architecture with 256 CUDA cores delivering over 1 Teraflops of performance.  It is a very good system for embedded deep learning, computer vision, graphics, and of course ALPR.
  • Jetson TX2 is the fastest and most power-efficient embedded AI computing device. This is one of the best supercomputer modules you can get that brings true AI computing at the edge with a power consumption of only 7.5 Watt.  It is built around an NVIDIA Pascal™-family GPU and loaded with 8GB of memory and 59.7GB/s of memory bandwidth. It has a lot of standard hardware interfaces that makes it easy to integrate into a wide range of products and form factors.
  • Jetson Xavier NX is very efficient in running modern AI workloads and consumes as little as 10 watts of power.  It has a compact form factor smaller than a credit card. Jetson Xavier runs modern neural networks in parallel and process data from multiple high-resolution sensors creating a platform for embedded and edge computing devices that demand increased performance but are constrained by size, weight, and power budgets.
  • Jetson AGX Xavier is the most powerful embedded system-on-module with respect to other Jetson modules and has the capability of a desktop GPU workstation while operating with just 10 Watts of power.  Jetson AGX Xavier makes neural networks easy to train and implement quickly and also opens the door for many AI-powered autonomous machines to thrive.  It has high computing performance with 32 TeraOPS.

Summary of Jetson devices. Prices are for USA market in USD as of April 27, 2020. Source: Plate Recognizer, Nvidia website.

What is Plate Recognizer ALPR?

Our Automatic License Plate Recognition software takes images as input and decodes the license plates in the image.  We can include the vehicle’s make, model, type and color into the decoded output as well.  

Our ALPR engine has been optimized to handle real-world scenarios like blurry, dark or low-resolution images.  Our software can even handle plates at different angles, when the vehicle is driving fast, or when the plate contains two rows of text. 

Optimizing ALPR on Jetson Devices

Since most Jetson boards comes pre-flashed with a Linux OS environment, it is easy to deploy and run Plate Recognizer ALPR.  

The high specs also enhance the performance with RAM of up to 4 GB in some of the Jetson boards and ARM based processor with high speed.  In addition, the Jetson devices have enough internal storage capacity to store more than 2 million images directly on the device.  Jetson devices also boasts of high connectivity with some of the boards supporting 4G Lte and GPIO pins. Jetson boards are generally very power-efficient with some working perfectly on 10W of power.

With the above specs, we have been able to optimize ALPR with an inference speed of 250 ms on the Jetson Nano device, with much faster ALPR inference speeds on the higher-power TX1, TX2 and AGX Xavier devices.

Installing ALPR on Jetson Devices

Plate Recognizer On-Premise SDK supports the following Jetson devices:

  • Nvidia Jetson Nano
  • Nvidia Jetson TX1
  • Nvidia Jetson TX2
  • Nvidia Jetson Xavier NX
  • Nvidia Jetson AGX Xavier

Setting up Plate Recognizer ALPR on Jetson devices is easy.  Just follow these steps: ALPR on Nvidia Jetson.

Plate Recognizer Runs on Jetson Devices. Now What?

We are excited that Nvidia has brought a new set of possibilities to deep learning.   We have seen customers use Jetson devices on a variety of ALPR use-cases, including:

  • Parking Access.  Incorporate ALPR with the Jetson device so that when the detected vehicle plate is on the whitelist, the system can automatically open the parking barrier gate.
  • Highway Monitoring.   Use a high-powered Jetson unit to process license plates in real time as vehicles are detected on the road.
  • Toll Management.  Utilize a medium-powered Jetson device to run ALPR on all the toll stalls within a toll plaza.  

What has been your experience with utilizing our ALPR On-Premise SDK with a Jetson device?  Let us know!

Privacy Settings
We use cookies to enhance your experience while using our website. If you are using our Services via a browser you can restrict, block or remove cookies through your web browser settings. We also use content and scripts from third parties that may use tracking technologies. You can selectively provide your consent below to allow such third party embeds. For complete information about the cookies we use, data we collect and how we process them, please check our Privacy Policy
Consent to display content from Youtube
Consent to display content from Vimeo
Google Maps
Consent to display content from Google
Consent to display content from Spotify
Sound Cloud
Consent to display content from Sound