Cart

Added to cart

Go to cart

Your order qualifies for free 2-day shipping and easy returns

FLIR AVP

Model: SOM 4251537
product image

In Production
The AVP, a size, weight, and power (SWaP) optimized advanced video processor, provides best-in-class artificial intelligence performance for thermal infrared and visible camera perception systems. It incorporates the latest Qualcomm® QCS8550, the industry’s most advanced mobile processor system on chip (SoC) featuring up to 48 TOPS inference compute performance. The QCS8550 is part of Qualcomm’s Product Longevity Program, ensuring future-proof product stability.

The AVP is designed to efficiently run Teledyne FLIR Prism AI software providing detection, classification, and target tracking and Prism ISP algorithms including super resolution, image fusion, atmospheric turbulence removal, electronic stabilization, local-contrast enhancement, and noise reduction. Prism software runs efficiently on the AVP leveraging the available system on module (SoM) functions to provide critical edge AI capabilities for automotive, airborne, unmanned, counter-UAS (CUAS), perimeter security, and intelligence, surveillance, and reconnaissance (ISR) applications.
 


FLIR-AVP-PDP-Pillar1-Images-INDUSTRY-LEADING-SWaP.jpg

INDUSTRY-LEADING SIZE, WEIGHT, AND POWER (SWAP) CONSUMPTION

The QCS8550 is fabricated on the advanced 4nm node

  • 40.27 x 33.41 mm
  • 6 W maximum and 2.5 W typical power consumption
  • 5 grams (without shields)
  • 4nm fabrication node enables low power
    and simplifies thermal management
 

FUTURE PROOF YOUR DESIGN

Utilize the highest-performance SoC in the industry

  • 48 TOPS INT 8 AI processor
  • 3.2 TOPS on GPU (float 32)
  • 8 CPU cores
  • 6 MIPI camera interface (5 concurrent)
  • USB camera interface


FLIR-AVP-PDP-Pillar2-Images-IFUTURE-PROOF-YOUR-DESIGN.jpg

 


FLIR-AVP-PDP-Pillar3-Images-DESIGNED-FOR-INTEGRATORS.jpg

DESIGNED FOR INTEGRATORS

Simplify development and reduce risk with Prism software and support

  • OS - Linux LE (r76)
  • Development support by qualified engineers
  • Sample applications included
  • Ongoing model development using realworld and proven synthetic data


Specifications

Hardware

Memory
16GB LPDDR5x
Operating Environment (OE)
Input Voltage: 3.6V Operating Temperature: -20 to +60 °C
Operating System (OS)
OS - Linux LE (r76)
Part Number
SOM 4251537
Devkit Part Number: 421-0100-00
Processor
Qualcomm QCS8550
64-bit Octa-Core
Application processor at 3.2 GHz (Gold+), 2.8GHz (4 x Gold), 2.0GHz (3 x Silver) Qualcomm Kryo CPU
Qualcomm Adreno GPU 740
Qualcomm Hexagon Tensor Processor (HTP) with Hexagon Vector eXtensions (HVX) and Hexagon Matrix eXtensions (HMX)
Qualcomm Secure Processing Unit for advanced secure use cases
Low Power AI (LPAI) subsystem with dedicated DSP and AI accelerator (eNPU) supporting always-on audio, sensors, contextual data streams, and Always-on camera.
Video
UHD video processing unit
AV1 decode
Native decode support for H.265 Main 10, H.265 Main, H.264 High, and VP9 profile 2
Native encode support for H.265 Main 10, H.265 Main, H.264 high formats

Mechanical

Dimensions
SOM Board: 40.27 x 33.41 mm LGA form factor
Weight
5 grams (without shield)

Storage

Storage
256GB UFS 3.1

Media gallery

Deploying AI Object Detection, Target Tracking, and Computational Imaging on Embedded Processors
Maximize Perception with Prism ISP and AI Tools for Enhanced Thermal Imaging | SPIE DCS 2024

Resources & Support

News articles and white papers written by Teledyne FLIR OEM experts to help you integrate your Teledyne FLIR OEM products.

For technical questions, find answers by visiting our FAQ at: flir.custhelp.com

Whitepaper

Deploying AI Object Detection, Target Tracking and Computational Imaging Algorithms on Embedded Processors

Learn more

Assets below are shared by all models in this product family.

FLIR AVP - Datasheet

Datasheet

Download

Camera Selector Tool

Compare all our infrared camera core models in one place.

Visit the Support Center for help with your Teledyne FLIR OEM products or submit a support ticket, by visiting our Technical Support Center.

Related Products