Focal-Plane Sensor-Processor Chips
Focal-Plane Sensor-Processor Chips Edited by: Ákos Zarándy Focal-plane sensor-processor imager devices are sensor arrays and processor arrays embedded in each other on the same silicon chip. This close coupling enables ultra-fast processing even on tiny,
- PDF / 11,852,698 Bytes
- 310 Pages / 439 x 666 pts Page_size
- 79 Downloads / 296 Views
´ Akos Zar´andy Editor
Focal-Plane Sensor-Processor Chips
123
Editor ´ Akos Zar´andy MTA Budapest Computer & Automation Research Institute PO Box 63, Budapest Hungary [email protected]
ISBN 978-1-4419-6474-8 e-ISBN 978-1-4419-6475-5 DOI 10.1007/978-1-4419-6475-5 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2011921932 c Springer Science+Business Media, LLC 2011 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper Springer is part of Springer Science+Business Media (www.springer.com)
Preface
Vision is our most important sensory system. It receives the largest amount of data and at the same time the most complex signal composition. Our eye and brain can put together images from photons arriving from different angles, with different densities and different wavelengths. One third of the human brain is devoted to the processing of the visual information. The processing starts already in the retina. It is responsible for adaptation and early image processing. Along the preprocessing phases, it reaches 1:100 data reduction. It is not surprising that the efficiency of the retina stimulated many research engineers to build vision chips, which mimic this amazing performance. The common properties of these sensor-processor chips are that they can perform both data capturing and processing. The story of the focal-plane sensor-processor (FPSP) devices started about 20 years ago, when Carver Mead built his famous silicon retina [1], which was capable of performing local adaptation. This initialized a line of different FPSP chips in the 1990s [2–6]. These chips were implemented using analog VLSI technology. The main motivation for the analog implementation was twofold. On the one hand, in the 1990s the silicon area of the basic computational elements needed for image processing (adder, multiplier, storage elements with 6- to 8-bit accuracy) was 5 times smaller in the analog domain than in the digital. On the other hand, the combination of the wide spreading CMOS sensors technology and the analog processing elements led to very efficient circuits, because no analog-to-digital conversion was needed. These chips contained 400–4,000 processing elements and could perform 10,000 FPS image capturing and processing real-time providing 10–100 times more computational power than a PC at that time. A summ
Data Loading...
