Wednesday, 29 June 2011

Trip 10 - Filter confirmation

The tenth trip to Delamere is really about confirming that everything we've put in place so far provides optimal performance at both wavelengths before we conduct our week long, intensive visit from 4th - 8th July. This visit is also an excellent opportunity to test out the new mount that has been custom made by the University Mechanical Workshop.
SALCA on new mount, which allows for levelling at all four corners.
Despite a late start and some decidedly ominous cloud cover, the scans all run smoothly and we have managed to collect some really good results. The coarser scans allow a bit more flexibility and we are hoping to try to fit in 8-10 sites a day next week. There is little change in the canopies now, as further leaf growth is more concentrated on bulking out the canopy as opposed to expanding the cover.
SALCA at Plot 3
The undergrowth at plots 2 and 4 continues to grow with fruit starting to appear on some of the brambles in plot 2 and the bracken at plot 4 exceeding 2 metres in some areas (this undergrowth in particular can be seen on the scans that we have produced).
SALCA at Plot 4 - in amongst the bracken
The resulting fisheye graph from Plot 4, which shows the influence of the undergrowth around the edge of the image.

Next week, an intensive study of the site will be conducted in an attempt to collect as much information over as short a time as possible. Consequently, alongside SALCA, we will also be using a Riegl 210z scanner as well as conducting various manual data collection methods including tree surveys and Diameter Breast Heights.

We are praying that the rain holds off.

Tuesday, 14 June 2011

Trip 9: More filter tests

Grabbing the only really nice day of the week, we have managed to complete 6 lower resolution scans today (two in each plot). As we have discovered that the 1040 laser is weaker than the 1550, different optical filters have been used: ND0.6 for the 1040 and ND 1.6 for the 1550.

SALCA at plot 2

Focusing on the Gap Fraction Analysis, there are a couple of points of interest:

1) SALCAs measurement of the Gap Fraction under-estimates the gaps compared to the hemispherical photography: This follow previous work completed at the University of Salford and arises due to the fact that hits and misses are only currently counted as either full hits or full misses - there is no allowance for partial or minimal hits.
Plot 4, where the ferns have reached head height and are showing up on our scans

2) At higher zenith bands (closer to the ground) we have found that the Gap fraction appears to increase, where the hemispherical photography analysis does not. We believe that this is a result of the range of the scanner only being able to pick up the nearer targets, whereas the hemispherical photos are guided by the number of photons hitting the lens.
Gap fractions for plot 4: Red lines indicate the 1040 scans and Blue lines show 1550 scans, with different noise thresholds measured of reach. The green line is the Gap Fraction from the hemispherical photographs

These experiments are to compare against the experiments we conducted last week to find an acceptable level of repeatability. There have been some issues with the current mount that is used to hold the filters and this is definitely something that needs work, but our results largely follow what we expected. Another trip will be needed to ensure that our results are acceptable. In the meantime, further tests will be run back at the university.

Oliver Gunawan
14th June 2011

Wednesday, 1 June 2011

Trip 8: Delamere Filter tests

The development of the canopies at all three plots is starting to slow down as they fill out. Therefore, it is not necessary for our research to continue on a weekly basis and we are now able to focus on fine tuning SALCA's recordings.

video

Today, we are experimenting with different filters using the hemispherical scan mode at plot three, which has been the latest to develop. In previous research, we have found that different filters allow different levels of information through to SALCA's recorder, with the 1040 returning noticeably weaker signals than the 1550, which itself produces some saturated returns.

The white panel pictured has a known reflectance
We currently use an ND1 filter, but will also be using ND0.6, ND1.2 and ND1.6 (larger numbers indicate stronger filters).
A selection of optical filters
We are also experimenting with a lower resolution scan, to see how well the trade off stands between data volume and speed of collection. If we find the results compare favourably with a full resolution scan, we could potentially complete many more scans in a single day.

No chair this week, but still no escape from marking

1st June 2011
Oliver Gunawan