Date of Award

January 2016

Document Type

Thesis

Degree Name

Master of Science (MS)

Department

Mechanical Engineering

First Advisor

William Semke

Abstract

The purpose of this research is to enhance visualization of warm-blooded animals and analyze the vegetation in which they are located using a camera system mounted on Unmanned Aircraft System (UAS). The results are coherently displayed in a single image at the same spatial location so that biologists will have accurate animal counts along with vegetation conditions. Application of aerial imagery has been used to analyze the vegetation health and determining the number of the animals by the wildlife service. However, a major obstacle of the research is to combine the two imaging systems to obtain the same spatial image with enhanced visualization. The two camera systems used in this research are the Tetracam ADC lite multispectral and the FLIR Photon 320 infrared camera. These two camera systems each have a different lens, field of view and sensor array size. The project involves the alignment of the two cameras to pixel level for the spectral image analysis. The spectral image analysis provides both vegetation information, such as Normalized Difference Vegetation Index (NDVI), along with enhanced visualization of warm-blooded targets. The system was miniaturized for the standalone payload for aerospace applications including UAS.

The FLIR Photon 320 was used to capture the infrared image and a Tetracam ADC lite multispectral camera was used to capture the near infrared, red and green spectral band images. A laboratory experimental setup was designed to mechanically align the two camera systems to get close identical spatial imagery. Spatial registration of the two images was performed using reverse image warping method by finding affine transformation matrix using point correspondences. Both camera systems were calibrated using Camera Calibration Toolbox for Matlab to reduce any distortion due to the lenses. A single board computer is used to capture and store the image data from FLIR Photon 320 infrared camera while the Tetracam image data is stored internally on board. The image capture time was set by continuous timed delay triggering within the Tetracam camera. The single board computer follows the Tetracam signals and matches the FLIR Photon 320 still image acquisition time with the Tetracam ADC lite.

Once the images are captured and stored by the camera systems, the files are downloaded and image processing is conducted. The data was analyzed to calculate the NDVI to observe the plant health. The infrared spectral band was used to identify the warm-blooded animals. In addition, various false color combinations of spectral bands and normalized difference ratios are processed to observe the visual enhancement capabilities on the vegetation and warm-blooded animal. It was determined that detected warm-blooded animal in the infrared band registered on top of NDVI image to show the vegetation health in a single image produced effective results. The combined image data from the FLIR Photon 320 and Tetracam ADC lite produced meaningful vegetation and animal information in the single image. This enhanced the capacity to identify and count animals while simultaneously characterize the vegetation environment, which is highly desired in ecosystem studies

Download

Share

COinS