Custom Image Processing

Nova has developed an extensive library of powerful hardware based image processing algorithms.  Some of these algorithms have been developed to perform spectral and atmospheric compensation while others have been designed to calculate real-time pixel-based FFT or PSDs.  Whatever your image processing algorithm, Nova’s capable team is ready to help you transition your algorithm model into a high-speed hardware implementation.

Example 1: Kernel-based edge detection

n x n convolution processing applied to real-time imagery produces an edge detection result.  A wide variety of spatial and temporal convolution processing may be performed to enhance target detection and increase target detection performance.                                         (MIG image courtesy AFRL/MNGI)

Example 2: Hypertemporal Target Identification*

Power spectral density (PSD) processing may be applied to specific target regions resulting in PSDs for individual pixels.  Long-range target detection and potential typing may be performed even on single-pixel objects.      * Joseph A. Kristl, J.Q. Peterson, M.F. Larsen, F.O. Clark, "Considerations for ELDT Hypertemporal Algorithms", 2003 Military Sensing Symposia, Monterey, CA, Jan 2003.

Example 3: Target Tracking Applications

Intensity- and motion-based target tracking algorithms have been developed for use in the Nova "Mini Image Processor".  Real-time tracking has been demonstrated using VASI    sensors and real-time spatial reconfiguration of the FPA in response to processor commands.

TM

Imaging the Future

Contact

About Us

Publications

SBIR Contracts

Amber Camera Service

Products and Services