|
Abstract:
|
The process of clustering and discriminate analysis on digitized imagery can be enhanced via vector augmentation to increase accuracy of the pixel discrimination into distinct classes. Our paper discusses an improved method of vector augmentation that permits greater accuracy when one seeks to identify specialized entities, like rivers, roads, etc. To obtain the digital imagery, an inexpensive Wi-Fi controlled vehicle was mounted with a digital camera and web-camera. In place of a handheld remote controller, a payload comprised of a miniaturized computer (the OQO, a 14 ounce computer) was added and used to do real-time processing. Using remote control of the OQO from a laptop, via a Wi-Fi/Bluetooth connection, control of operations was accomplished aboard the vehicle, and real-time multivariate analyses were conducted on digital images acquired by a high-resolution camera.
|
