A reconfigurable real-time compressive-sampling camera for biological applications

Many applications in biology, such as long-term functional imaging of neural and cardiac systems, require continuous high-speed imaging. This is typically not possible, however, using commercially available systems. The frame rate and the recording time of high-speed cameras are limited by the digit...

Full description

Saved in:
Bibliographic Details
Published in:PloS one 2011-10, Vol.6 (10), p.e26306-e26306
Main Authors: Fu, Bo, Pitter, Mark C, Russell, Noah A
Format: Article
Language:English
Subjects:
Acoustics
Animals
Automation
Bandwidth
Bandwidths
Biology
Biophysics
Blood cells
Cameras
CMOS
Continuity (mathematics)
Daphnia - physiology
Data Compression - methods
Digitization
Drosophila
Engineering
Heart - physiology
Heart diseases
High speed cameras
Image acquisition
Image processing
Insects
International conferences
Kinematics
Light
Molecular Imaging - instrumentation
Molecular Imaging - methods
Moving targets
Neurosciences
Physics
Real time
Science
Sensors
Signal processing
Target recognition
Time compression
Time Factors
Wavelet transforms
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Many applications in biology, such as long-term functional imaging of neural and cardiac systems, require continuous high-speed imaging. This is typically not possible, however, using commercially available systems. The frame rate and the recording time of high-speed cameras are limited by the digitization rate and the capacity of on-camera memory. Further restrictions are often imposed by the limited bandwidth of the data link to the host computer. Even if the system bandwidth is not a limiting factor, continuous high-speed acquisition results in very large volumes of data that are difficult to handle, particularly when real-time analysis is required. In response to this issue many cameras allow a predetermined, rectangular region of interest (ROI) to be sampled, however this approach lacks flexibility and is blind to the image region outside of the ROI. We have addressed this problem by building a camera system using a randomly-addressable CMOS sensor. The camera has a low bandwidth, but is able to capture continuous high-speed images of an arbitrarily defined ROI, using most of the available bandwidth, while simultaneously acquiring low-speed, full frame images using the remaining bandwidth. In addition, the camera is able to use the full-frame information to recalculate the positions of targets and update the high-speed ROIs without interrupting acquisition. In this way the camera is capable of imaging moving targets at high-speed while simultaneously imaging the whole frame at a lower speed. We have used this camera system to monitor the heartbeat and blood cell flow of a water flea (Daphnia) at frame rates in excess of 1500 fps.
ISSN:1932-6203
1932-6203
DOI:10.1371/journal.pone.0026306