|
The Computer Vision Group conducts research and development of algorithms and software for automatically interpreting data from a variety of imaging sensors in any region of the spectrum, particularly cameras and ladar/lidar, but also including radar. The group's application focus includes 3-D perception, position estimation, motion estimation, and terrain analysis for autonomous navigation of planetary rovers, safe and precise landing of landers, and landmark recognition for orbiters. The group also develops algorithms to automate the interpretation of imagery from science instruments, with the implementation running either on Earth or onboard the spacecraft. In addition, the group addresses applications for a variety of sponsors outside of NASA, including autonomous navigation of unmanned ground, air, and sea vehicles, object recognition, and video reconnaissance. Finally, in collaboration with other groups at JPL, the group develops concepts and prototypes for custom hardware implementations of smart imager chips, smart cameras, compact ladars, and compact, low power, high performance image processing systems, such as based on field-programmable gate arrays (FPGA).
|
