RFEL signal processing iq : fpga expertise

RFEL

Video Processing

IP Cores

NICE

Upgrade your surveillance, security or situation awareness system to reveal unseen detail and improve the appearance of your video, using RFEL’s Non-linear Image Contrast Enhancement (NICE™) technology. NICE is a hybrid FGPA and software solution, optimised for System-on-Chip (SoC) devices such as the Xilinx® Zynq®-7000.

Non-linear Image Contrast Enhancement Examples

Features

  • Maintains operation as scene visibility deteriorates.
  • Both low-light and daylight enhancement for a versatile solution.
  • Low latency of <1ms for human-in-the-loop applications.
  • Rapid adaption (<100ms) to changes in illumination.
  • Support of full HD at high frame rates.
  • The NICE module can be combined with other RFEL IP modules in a single hardware device.

Applications

The versatility and low resource usage of NICE means that it can provide product differentiation for a huge range of video applications. NICE can dramatically optimise the contrast of the displayed video, irrespective of changes of illumination or atmospheric conditions.

Good image contrast is critical for human perception as it aids the interpretation of images and reduces reaction times, thereby increasing the effectiveness of video surveillance and border control.
NICE also offers improved situation awareness, which is important for the control of vehicles and in combat situations, where a rapidly changing degraded visual environment (DVE) can threaten security and safety. Generic inspection and screening tasks, such as those needed for manufacturing or medical applications, may also benefit from NICE’s high-speed enhancement techniques.

NICE provides a cost eff ective solution to improving both the quality and operational robustness of a vision system. Human visual perception is primarily based on intensity and so the correct management of scene brightness and contrast can make profound visual diff erences to the quality of your video images. NICE is a relatively simple processing operation, but its benefi ts are far from basic.

  •   
  •   
  •   
  •   

Please enter the word you see in the image below:


 

 

Non-linear Image Contrast Enhancement Block Diagram

The NICE algorithm is cleverly implemented to optimally utilise the high-speed, low power FPGA processing resources, but also the embedded CPU cores, of state of the art system-on-chip (SoC) devices, such as the Xilinx Zynq-7000. The result is a highly efficient and flexible embedded processing engine that adds competitive advantage to your video capability. Engineering consultancy services are available to assist in integration of the IP, or RFEL sales can be contacted to discuss how NICE can be supplied as a plug-in system-on-module (SoM), or as a fully integrated sub-system for turn-key integration.

Sub-item

Description

Performance Automatically corrects gain and improves low-contrast details to dramatically enhance video
Algorithm Contrast limited histogram equalisation for daylight operation and a proprietary technique for optimum low-light operation
Operating Modes Low-light, daylight or automatically selected
Interfaces AXI4 streaming and memory interfaces
Input Format(1) Up to 12 bit mono, 16 bit YCbCr 422, 24 bit YCbCr 444 and 24 bit RGB
Output Image Format

Identical to the input

(Contact RFEL for conversion solutions)

Resolutions Up to and including 1920x1080
Frame Rates Up to and including 60 fps, progressive
Latency <1ms
Temporal Adaptation <100ms
Clock Frequency(2) 200MHz
Supported Devices

Xilinx Zynq™-7000 / Artix™-7/Kintex™-7

Reference Design RFEL HALO™ RPDS based pre-integrated Quick Start hardware dev. kit
Delivery Xilinx's ISE net-list or Vivado IP builder

(1) Delivered formats factory configured
(2) Xilinx Kintex targets

 

Ordering Information

Item

Part Number

NICE Xilinx Zynq - 7020 IP 10-05-ZA
NICE Xilinz Zynq - 7030/45 IP 10-05-ZK
NICE System Consultancy Contact RFEL Sales

Case Study FPGA resource usgae on Zynq™-7020

Item

Usage

Registers 663 (<1% of resource)
LUTs 454 (<1% of resource)
Block RAMs 4 (3% of resource)
DSP48Es 0 (0% of resource)

Latest News