ADAS Vision Analytics Toolbox (AVAT)

ADAS Vision Analytics Toolbox (AVAT) is designed to be a software toolbox for system integrators working on Vision based ADAS applications. AVAT is a set of software modules that integrate primitives from ADI Image Processing Toolbox and provides solutions to relatively complex problems like Circle detection, Number Recognition etc. This is at a level that is lower than a full scale application and higher than image/video processing primitives. These modules can be used by ADAS applications like Traffic Sign Recognition or Lane Departure Warning system.

ADAS Vision Analytics Toolbox (AVAT)

Product Description

ADAS Vision Analytics Toolbox (AVAT) is designed to be a software toolbox for system integrators working on Vision based ADAS applications. AVAT is a set of software modules that integrate primitives from ADI Image Processing Toolbox and provides solutions to relatively complex problems like Circle detection, Number Recognition etc. This is at a level that is lower than a full scale application and higher than image/video processing primitives. These modules can be used by ADAS applications like Traffic Sign Recognition or Lane Departure Warning system.



Features


Functions


Performance Metrics

MIPS summary:
Input Vector Module Code RAM (KiB) Data RAM (KiB) Constant Data Tables (KiB) AverageCycles/Pel MIPS
avat_demo_inp_12.pgm Canny Edge Detector 9.79 125.27 0.007 16.57 153
avat_demo_inp_12.pgm Sobel Edge Detector 6.05 300.16 0.33 14.49 134
avat_demo_inp_12.pgm Hough Transform Circle 1.01 300.16 0.33 14.49 134
avat_demo_inp_12.pgm Classifier 1.01 0.048 0 278970 8
avat_demo_inp_12.pgm Feature Generation 3.21 2.09 0.08 17.74 163
avat_demo_inp_12.pgm Shape Detector 4.60 300.20 1.07 16.134 149

  • Performance is measured on ADSP-BF561 processor silicon revision 0.5
  • Data Cache and Instruction Cache are enabled
  • "Data RAM" for one instance, includes Scratch, Instance/State memory
  • MDMA:One channel Memory DMA is used
  • Image width=640 and height=480, with MDMA disabled
  • Edge Detector: Percentage of edge pixels is 10%.
  • Hough Transform: Min Radius = 64, Max Radius = 74, Scale = 2 and accumulation threshold = 250.
  • Classifier: Number of trees = 250, only one ROI per image. Number of classes = 16.
  • Feature Generation: roi_width=130, roi_height=130, number of vertical sub blocks=2, number of horizontal sub blocks=3.
  • For the classifier module the cycles value specified is the total number of cycles taken per classification.
  • MIPS is measured as ((Avg cycles/pel) * (image_width) * (image_height) * (frame_rate) / 10^6) image_width=640, image_height=480 and frame_rate = 30.
  • For classification MIPS is calculated with an assumption of 30 classifications per second.
  • 1 KiB = 1024 Bytes.

Applications


Requirements


Availability and Licensing

Each module supports the Analog Devices, Inc. (ADI) Blackfin Processor family and is a licensed product that is available in object code format. Recipients must sign a license agreement with ADI prior to being shipped the modules identified in the license agreement.

Contact your ADI Sales Rep to request this code. If you need to find a Sales Rep in your area, please visit the Sales & Distributor Map/Listing.


沪ICP备09046653号
content here.
content here.

Review this Product

Close