Design and Implementation of Vision Signal Processor with OpenCV Library Support
Date Issued
2008
Date
2008
Author(s)
Lin, Chia-Hua
Abstract
An efficient architecture design of vision signal processor with OpenCV library supports presented in this thesis. It is a 64GOPS, 86.5mW vision processor which is implementedn a 2.75mm?2.75mm die in a UMC 90nm Logic&Mixed-Mode 1P9M Low-Krocess.n the resent decades, video analysis technology plays more and more important rolen many vision applications, such as surveillance system, healthcare, intelligent vehicleystem and so on. It is believed that intelligent video analysis technology will must behe trends of development.owadays, Intel OpenCV library is popular in the research domain and creates lot’sf successful applications. Our hardware can provide one to one function mapping fromC OpenCV library to embedded system. With this hardware, the implementation timean be saved and speed up the product become available in the market.ne of the most frequently used operations in image recognitions morphologicalrocessing, which is often adopted for pre-processing of various applications based onmage recognition. Experimental result shows our proposed architecture performs highomputation throughput and low power consumption. It can process 1024?768 8-bitray level image with more than 200 frames per second and the area is quite smallompared to state-of-the-arts.obust and rapid object detection is the other challenge in the field of computerision. A object detection algorithm proposed by Viola and Joses is implemented inur design. We optimize the algorithm to our hardware with low performance drop.etection rate for CMU+MIT test database which consists of 119 images with 513abeled frontal faces is 87.3%. The processing speed is 10 frames per second with20?240 8-bit gray level image
Subjects
vision
processor
opencv
face detection
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-97-R95943021-1.pdf
Size
23.32 KB
Format
Adobe PDF
Checksum
(MD5):70f4b56f93ac190ecea357c172887953