Integrated motion vision sensor

Details Integrated motion vision sensor Integrated motion vision sensor

1998-07-09

2001-06-26

Shah, Kamini (Department: 2857)

Data processing: measuring, calibrating, or testing

Measurement system

Orientation or position

C702S152000, C702S187000, C382S209000

Reexamination Certificate

active

06253161

ABSTRACT:

FIELD OF THE INVENTION
The present invention relates to a motion vision sensor. The invention relates further to a motion vision sensor with self-signaling pixels.
1. Introduction
Image motion computation could benefit applications such as self guided vehicles, mobile robotics, smart vision systems and intelligent cameras. On the other hand, motion computation algorithms have large computational overheads, requiring specific architectures for real-time operation.
REFERENCES
The following references are of background interest to the present invention, and will be referred to hereinafter by the reference numeral in square brackets:
[1]
Computational Sensors, Report from DARPA Workshop,
T. Kanade and R. Bajcsy, Eds,. University of Pennsylvania, May 11-12, 1993.
[2] R. Sarpeshkar, W. Bair, C. Koch, “Vision Motion Computation in Analog VLSI using pulses,” in
Neural Information Processing Systems
5, S. Hanson, J. Cowan, C. Giles, (eds.), Morgan Kaufman: San Mateo, pp. 781-788, 1993.
[3] R. Etienne-Cummings, S. Fernando, N. Takahashi, V. Shtonov, J. Van der Spiegel, P. Mueller, “A New Temporal Domain Optical Flow Measurement Technique for Focal Plane VLSI Implementation,” in
Proc. Computer Architectures for Machine Perception
, pp. 241-250, 1993.
[4] M. Arias-Estrada, M. Tremblay, D. Poussart, “A Focal Plane Architecture for Motion Computation”, Journal of Real-Time Imaging, Special Issue on Special-Purpose Architectures for Real-Time Imaging, in press.
[5] T. Delbrück and C. Mead, “Phototransduction by Continuous-Time, Adaptive, Logarithmic Photoreceptor Circuits,” Tech. Rep., California Institute of Technology, Computation and Neural Systems Program, CNS Memorandum 30, Pasadena, Calif. 91125, 1994.
[6] C. P. Chong, C. A. T. Salama, K. C. Smith. “Image motion detection using analog VLSI.” IEEE J. Solid-State Circuits, Vol. 27, No. 1, pp 93-96.
[7] J. Lazzaro, J. Wawrzynek, M. Mahowald, M. Sivilotti, D. Gillespie, “Silicon auditory processors as computer peripherals,” IEEE Transactions On Neural Networks, Vol. 4, No. 3, pp. 523-527, May 1993.
[8] T. Delbrück, “Investigations of Analog VLSI Phototransduction and Visual Motion Processing,” Ph.D. Thesis, Dept. of Computation and Neural Systems, California Institute of Technology, Pasadena, Calif., 91125, 1993.
[9] T. Delbruck, C. Mead “Adaptive photoreceptor including adaptive element for long-time-constant continuous adaptation with low offset and insensitivity to light.” U.S. Pat. No. 5,376,813
[10] T. Delbruck, C. Mead, “Subthreshold MOS circuits for correlating analog input voltages.” U.S. Pat. No. 5,099,156
[11] M. Mahowald, M. Sivilotti, “Apparatus for carrying out asynchronous communication among integrated circuits.” U.S. Pat. No. 5,404,556
Computational sensors [1] integrate sensing and processing on a VLSI chip providing a small, low-power and real-time computing front-end. Analog processing is often used because it interfaces nicely to the real world, it requires little silicon area and it consumes low power. Analog computation low precision is not essential for motion computation. In this document, we describe a motion computation imager which detects motion at the pixel level using analog processing techniques and communicates data through a digital channel. Velocity is computed in the digital domain during off-chip communication. The implemented architecture overcome some of the limitations and difficulties encountered on other focal plane architectures for motion computation [2, 3, 4] and overpasses largely traditional systems based on CCD imagers and a CPU.
This document is organized as follows: first, a description of the motion computation algorithm is given. Then an architecture overview is described from the pixel level to the communication protocol and the velocity computation process. The next section details the VLSI implementation, describing the pixel, the communication circuitry and the system organization. Finally, some preliminary results of the motion sensor architecture are presented.
SUMMARY OF THE INVENTION
It is a first object of the invention to provide a motion sensor which detects moving features and stores the time stamps for later computation of velocity. It is a second object of the invention to use an address-event protocol for the motion computation paradigm. It is a third object of the invention to compute the velocity by using the focal plane for detecting moving features, using an address-event protocol combined to a time-stamp protocol to facilitate further data manipulation, and a direct one-pass velocity algorithm. It is a fourth object of the invention to provide a relatively low complexity pixel, providing practical fabrication of high density arrays with standard CMOS technologies. It is a fifth object of the invention to provide a motion sensor in which velocity computation precision is programmable and dependent on an external digital timer instead of multiple analog RC constants in the focal plane, as is the case in previous approaches. It is a sixth object of the invention to provide a motion sensor in which analog VLSI computation techniques are combined with digital computation techniques. It is a seventh object of the invention to provide a motion sensor in which the architecture can be integrated with other focal plane functionalities for multifunctional vision sensors. It is an eighth object of the invention to provide a motion sensor having an SRAM cell with non-retriggerable input and asynchronous reset. It is a ninth object of the invention to provide a motion sensor in which a one pass algorithm is used to compute velocities based on the time-stamps. It is a tenth object of the invention to provide a motion sensor in which a simplified digital architecture is used to implement the one-pass algorithm which can be integrated in a sensor control module.
As will be appreciated, the invention can provide one or more of the following advantages:
Continuous velocity vector field extraction
High density array of pixels
High density array of motion vectors in real-time
Real-time operation
Minimize the use of transistors in analog VLSI susceptible to mismatch and process variation. Minimize the number of analog voltages to bias the aVLSI structures reducing routing complexity
Compact three-chip solution to the motion computation paradigm: focal plane sensor, RAM memory, digital interface
Potential low-cost fabrication
Low power consumption, robust operation.
To in part achieve the above exact objectives, the invention provides a system for detecting and determining the direction of motion of an object, an image of which object appears as an image formed by an array of pixels. The system has a motion sensor with self-signaling pixels for signaling detection of motion and transmitting information regarding that motion. It also has a digital module which receives the information regarding the detected motion, which module stores this information and after a fixed measurement time period, computes velocity vectors from said information received from a set of at least two pixels and then transmits the velocity vectors to a host computer for further use. The motion sensor and digital module thus combined allow for real-time continuous operation of the system.
In a further aspect of this invention, the motion sensor has an array with a plurality of self-signaling pixels. The system then computes velocity vectors with a set of four adjacent pixels so that the digital module computes a bi-dimensional velocity vector. The system thus computes a plurality of velocity vectors which are sent as a vector map to the host computer.
The invention also provides a method for detecting motion of an object and determining the direction and speed of that motion, the method having the steps of sensing motion of an object with a plurality of self-signaling pixels, which pixels make up an array of motion sensors. Signaling by each of the pixels as it detects motion. Sequencin

Affiliated with

Also associated with

Rating

Integrated motion vision sensor does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Integrated motion vision sensor, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Integrated motion vision sensor will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2536290