Project ideas

From this summer, I've put together a list of possible projects that could be tackled by ambitious (and relatively tech savvy) students for an MA or honor thesis. 

• Mapping and integrating vertical imaging of surfaces with horizontal imaging (I.e., "greg's" project). This is largely a matter of figuring out how to best integrate the different datasets. Greg focused on the "images" and found they couldn't be matched — which isn't a surprise given the way each program deals with them. But the data are different — the XYZ coordinates. Those should be integrate-able. The images are a separate matter.
• Detecting and mapping cave features with TIR, NIR and VIS imagery. Jeanette didn't integrate the multiple sources of imagery into a single classification so couldn't get a good product. Im not sure why she didn’t do that other than "not enough time." But this would be the way to go.
• Multispectral camera (V1.0) use in mapping vegetation. This would be a project that would have a student work on the existing MS camera to map vegetation in an area of interest (e.g., Palo Verde or somewhere else). The current camera is functional and Im working on it to reduce the weight. This way it could be flown with the quadcopters — which means that students can do their own data collect. They would need to be wiling to:  learn some python to process the data, learn to futz with the electronics, have an area interest. But it would be a good project that builds on what we have already and the work that the Montana State University folks did (as ours is a variant of their camera but with more bands). 
• Multispectral camera (V2.0) The next version of the MS camera is going to be Raspberry Pi based. It will build on the concepts of V1 but have a lot more onboard processing and much higher resolution (and be lighter). What I envision is a camera that has 1 Rpi for each band with a single Rpi acting as controller and data integrator.  The cameras will be the tiny 5.0 megapixel cameras. This would be a great MA project I think. The goal will be to improve on the original  design and make it more modular with better on-board processing (such as automatically producing NDVIs). The student would need to learn python, GDAL, linux and be willing to learn about how the electronics work (but its all fairly simple). 
• Thermal camera image integration with UAVs — We need someone to take on figuring out how to take thermal imagery with the camera and integrate this into a geospatially referenced mosaic. The project would be electronic, programming, UAVs, and could focus on studying groundwater discharge along the coast of PV (or some other handy area). 
• Thermal camera mapping of archaeological features. Measuring differences in temperature due to buried rock features has been shown to be a great way to detect archaeology. No one has done this with a platform as small and mobile as what we have. Someone needs to figure out the best procedure for doing this, design a good test study, map some existing features (could do this in Mississippi, rapa nui, guatemala). 
• Thermal camera detection of artifact composition – Paul Buck (DRI) has demonstrated that long wave sensing can be useful for doing sub pixel mapping of varying density of artifacts. Id like to see if we can use this approach in a micro artifact identification capacity. This would combine image processing (perhaps via ImageJ) and image classification (perhaps via eCognition) to count artifact classes automatically. We would use the thermal camera in a microscope arrangement… 
• Temper identification in prehistoric ceramics: similar to above but examining how TIR can be used to quantify the composition of ceramics from image analysis alone. 
• Image analysis for mapping surface archaeological features with X100 We would fly X100 over landscape to map the surface in terms of topography, archaeology, hydrology. We need someone to work on extracting archaeological distributions from the imagery (would need to figure out best sets of filters/wavelengths, techniques in eCognition and so on).
• Constructing the ideal app for doing aerial photograph connected with photoscan. Someone needs to figure out the timing, georeferencing and processing. This could be a commercial application if done well. The person would need to be willing to learn some programming (well, quite a bit of programming).