Saturday, 24 December 2011

Terrestrial laser scanner workshop

The Salford group along with University of Lecester are organising a Knowledge Exchange Workshop at the University of Leicester in April 2012. Details below:

Friday, 4 November 2011

Range and reflectance calibration measurement

We have been testing the range dependence of the SALCA returns in a series of indoor experiments. The objective is to derive an empirical range corrections that can be applied to the waveforms in order to move forward with quantitative analysis of the intensity information. Alberto has derived a field based method for the Riegl scanners and we will compare this approach with the results from the lab calibration.

Monday, 17 October 2011

SALCA back in business

October has seen some changes for the SALCA project with Oliver Gunawan moving from the SALCA research assistant position to take up a PhD in the School. Oliver will be working on spatial mapping of urban areas and will incorporate airborne and maybe terretrial laser scanning into the project.

Dr Alberto Ramirez finally gets his hands on SALCA
October also saw Alberto Ramirez complete his PhD research using our Riegl TLS system to characterize multi-temporal changes in forest stands at Delamere Forest. Alberto has now taken over from Oliver and is starting to run a series of new calibration tests following the return of SALCA from Halo Photonics. Some adjustments have been made to the power of the lasers in order to better balance the return intensities from the two wavelengths. A new filter mount has allowed us to dispense with the cardboard and sticky tape, and the pulse repetition has been modified in 'stare' mode.

Monday, 19 September 2011

SALCA goes to Halo and we go to Bournemouth!

SALCA was finally delivered back to Halo Photonics, 2 hours down the M6 to Worcester. A useful discussion with Guy Pearson ironed out the amends that need to be completed on SALCA, which include:

1) Repositioning of the laser beams or over-rotation of SALCA to eliminate a blind spot that we have discovered.
2) Tweaking of beam magnitude in order to use only 1 ND filter. This will remove the need to perform consequent scans for each wavelength and will therefore greatly increase the quality of our data collection.
3) Installation of a more powerful digitiser that we are hoping will help toward speed SALCA up, either by continuous rotation of the head (without pause for download), or by the increased memory capacity, which would enable one single download at the end of a scan.
4) Investigation into SALCA beam power loss over time.
5) A more permanent mount designed to the amended filter angle, which should allow us to use and change filters on SALCA more easily.

Meanwhile, myself, Mark and Rachel attended the Remote Sensing and Photogrammetry Society Conference 2011 last week in Bournemouth, where Mark and Rachel were both well received in presenting the progress we have made, both practically and scientifically with the SALCA project, paying particular attention to the fieldwork completed and also the progress made with our moisture loss experiments.

Provisionally, we are hoping to retrieve SALCA in early October, which means we will have time to run more calibration tests before taking the scanner back to Delamere to try to catch the leaf off process, which should provide us with more useful phenological data.

Oliver Gunawan
19th September 2011

Friday, 12 August 2011

Leaf Drying and Range Calibration

Before SALCA is taken back to Halo Photonics, we need to complete a couple of experiments in order to make better use of the data sets we have already collected and to see if future plans are viable. Consequently,I have completed a leaf drying test and also a close-range calibration experiment.
Leaf used in experiment

Sand paper, Spectralon panel and mounted leaf

The leaf experiment was carried out using a fresh, whole leaf, which was fitted into a pre-weighed mount and placed next to 2 panels of known reflectances. The three targets were then scanned repeatedly over the course of 3 days with the mounted leaf being weighed after each scan. The purpose of this experiment is to see if the ratio of the two wavelengths changes during the drying process (as theory indicates), which would then be a useful indicator of moisture levels within canopies.
Leaf drying experiment

The close-range experiment is important because we have found a fluctuation in intensities at ranges of 0.25 m - 7 m. This experiment was conducted using 2 panels of known reflectance. One panel was kept static and another was placed at different distances from SALCA before being scanned. 16 scan positions were measured up to 7m, which should provide useful information for characterising the fluctuation in order to correct any values within this range.
Range Calibration: Left hand panel is stationary. Right hand panel moves in between scans.

Oliver Gunawan
12th August 2011

Monday, 25 July 2011

Taking the Riegl 390 out

I am accompanied today by Alberto Ramirez to complete scans of some of the plots we visited at the beginning of July using the Riegl Z390 laser scanner. This is a commercial scanner, which has been used extensively by Alberto during his PhD research.
The Riegl Z390 at plot 2

As a commercial unit, the Riegl is much quicker and simpler to set up, however we can only extract the first returns from the collected datasets. We are aiming to gather comparative gap fraction data to use against SALCA and the hemispherical photography. 2 ND0.6 filters were attached side by side to the receiving lens of the scanner as previous research has found the intensities saturate, making further analysis difficult. Using a resolution of 0.1 degrees in both the elevation and azimuth, we complete 2 6 minute scans at each plot (the second scan being positioned perpendicular to the first to collect the full hemisphere).
Alberto with the Riegl at plot 3

Building on Alberto's work, we have managed a full 'point-based' approach to gap fraction analysis, where the beam either hits a target 100% or not at all. This produces an underestimation of the gap fraction (the gap appears smaller) because sometimes only part of the beam is interrupted. By weighting by intensity and using only the first return, we are now overestimating the gap fraction, as we are not considering the interception of further returns. The next step is to focus on the effects of additional returns in sequence.
Robo-baby being transported to plot 1

We manage a full day of research, scanning an unprecedented 8 sites across delamere, which should supply a good range of canopy compositions and densities:

Plot 1: Corsican Pine
Plot 2: Oak / Sweet Chestnut / Silver Birch
Plot 3: Sweet Chestnut
Plot 4: Corsican Pine
Plot 5: Larch
Plot 10: Open Corsican Pine
Plot 11: Dense Corsican Pine
Plot 14: Dense mixed Deciduous

Oliver Gunawan
25th July 2011

Friday, 8 July 2011

SALCA Field Trip. Final Day

Thursday 7th July was the last full day of experiments planned in Delamere. In order to get our SALCA scan list up to date, we opted for 1 more Coniferous stand and a larch stand. Alberto and Vishal joined us once again to help conduct field measurements.
Plot 5: Larch stand

As well as the manual field measurements mentioned in previous posts, we also take an opportunity to collect GPS positions of each plot using new Trimble devices acquired by the University of Salford.

Alberto and Vishal conducting field measurements at plot 11

The picture below shows an example of a tree infected with Red Band Needle Blight. We were informed by the Forestry Commission that this disease mainly affects Corsican Pines, although the cause is not yet known. For this reason, our final plot was by  a cross roads between two pine stands, one of which was more affected by the blight. We are hoping that differences in spectral properties may be picked up in SALCA, which could be useful in monitoring change over time of the disease in particularly badly affected areas.
Red Pine Needle Blight mainly affecting Corsican Pine

Alberto, Vishal, Oliver and Richard by plot 16

The final day completed, we celebrated with a fantastic Chinese meal in Hatchmere, not 5 minutes drive from the forest car park. Friday morning was all about getting packed up before taking one last hemispherical photo from plot 1 to complete the set. We have collected almost 50Gb of data and over 100 photographs covering 12 plots:

Plot 1: Coniferous (mainly scotts pine)
Plot 2: Mixed deciduous (sessile oak/sweet chestnut/birch)
Plot 3: Deciduous (sweet chestnut)
Plot 4: Coniferous (scotts pine/corsican pine)
Plot 5: Larch
Plot 10: Open coniferous (scotts pine)
Plot 11: Dense coniferous (corsican pine)
Plot 12: Mixed deciduous (perdunculate oak/sweet chestnut/birch)
Plot 13: Open deciduous (sweet chestnut picnic area)
Plot 14: Dense deciduous (mixture of 6 species)
Plot 15: Open deciduous (Beech)
Plot 16: Between 2 coniferous stands

Let the data processing commence!

Oliver Gunawan
8th July 2011

Wednesday, 6 July 2011

SALCA Field Trip day 3

We are joined today by Richard Armitage, Alberto Ramirez, Vishal Bandagula and Turkia Almoustafa from the University of Salford, who have been involved in previous laser scanning research at Delamere forest. Despite some seriously pessimistic weather forecasts on the breakfast news, the day ends up being quite nice and with only a couple of showers interrupting play, we have managed to get a lot done today.

In the morning, we completed an experiment that Richard Casey has put forward, which involves placing a sheet of material with a known reflectance a distance behind leafy and woody material at plot 12, a  mixed deciduous plot. This area was then scanned at high resolution using both wavelengths. When the results have been processed, we will be able to isolate intensities that have come from 100% leaf hits and 100% wood hits. Under the assumption that the reflectance of leafy and woody material does not vary, the intensities collected could give us information on what material the beam has hit and by introducing a linear scale, intensity values can show us the proportion of the beam that the target has covered, which will potentially increase the precision of our gap fraction analysis results.

Blanket experiment at plot 12

One of our hopes was to use a Reigl 210 scanner to compare our SALCA scan results at each of the plots. However, there has been a technical issue with the Reigl and despite our best efforts, we have not been able to overcome these in the field. We do aim to take the 210 out at some point in the near future, as the data we will collect will be useful in linking this work with previous research conducted by Alberto and Vishal.
Mark and Turkia collecting spectral refectances of leafy and woody material using the ASD

In addition to further manual field measurements completed in plot 4, our second scan is completed in an open beech plot (plot 15). This stand has large, old trees and is ideal for a range calibration experiment. As we have not tested how laser intensity changes with range, we place calibration cards with a known reflectance on 8 trees at increasing distances from SALCA. We will be able to compare the results from the intensities of these cards against a theoretical loss of power (1/range squared).
Oliver explaining the operation of SALCA to Richard Mostyn form the Forestry Commission

During setup at plot 15, there is a chance encounter with several members of the Forestry Commission. This proves to be a great opportunity for an impromptu demonstration of SALCA and the chance to explain how our research fits into future forest monitoring and management strategies. We also use this as a chance to highlight the benefits of an off-road pram for intensive forestry fieldwork.

Oliver Gunawan
6th July 2011

Tuesday, 5 July 2011

SALCA Field Trip day 2

The conditions outside are looking considerably less bright than yesterday, so we head into Chester to collect equipment for experiments we are planning this week.
SALCA at plot 13 (open deciduous)

As we have decided not to scan this morning, we decide to start the manual field data collection, of plots 2 and 3, which include GPS positioning of the plots and tree measurements such as height (using a clinometer), diameter breast height and species composition, which could be important considering possibile variability of bark and leaf reflectances.
Plot 14 (Dense deciduous)

The afternoon  brightens up and two deciduous sites are chosen for scanning. The first (plot 13) is near a picnic area and has been chosen as it is a relatively open stand, with large, but dispersed sweet chestnut trees. The second stand chosen (plot 14) is a much denser stand composed of a greater variety of deciduous trees including sweet chestnut, birch, rowan, oak and wild cherry.
Moisture experiment setup. As SALCA is an active laser sensor, scanning in the dark is fine!

Whilst at plot 14, we also collect a selection of dead and live sweet chestnut leaves for further experiments back at the cottage. The wavelengths used by SALCA have the potential to provide a useful index of vegetation moisture content, as high leaf moisture will result in high absorption at 1550 nm. In constrast, the 1040 nm laser will mainly be sensitive to structural characteristics of the leaf. The experiments are the first test of the sensitivity of SALCA to vegetation health and leaf moisture content. A live leaf, a dead leaf from the forest litter layer and a partially oven-dried leaf are scanned at high resolution in both wavelengths to compare the intensity of the return signal. A ratio of the two wavelengths could provide an early indicator and useful measure of vegetation stress, drought and disease in woodland as well as estimates of forest fuel moisture content, a significant factor in determining fire risk. Early results look promising and future experiments will look to quantify the relationships between moisture content and SALCA intensity - check back soon for more results!

Oliver Gunawan & Rachel Gaulton
5th July 2011

Monday, 4 July 2011

SALCA Field Trip day 1

Today is all about making sure everyone gets to the site, although we make the most of the hot sunshine, by getting our first scan underway early.

In the morning, we selected several additional plots in order to get a wide variety of different species and densities of stands in both coniferous and deciduous areas. We now have a target of 10 sites including the three core plots that we have already covered.
Rachel, Mark and Richard with SALCA at plot 1
Today, Mark and I are joined by Rachel Gaulton from Newcastle University, who worked with SALCA for a year before I started and has been heavily involved in the calibration and creation of processing methodologies for SALCA and Richard Casey from University College London, who has started a PhD focusing on computer simulation of multi-waveform terrestrial lidar data and who is using some data from SALCA in his research.

Richard, Rachel and Oliver at Plot 11:  A dense Corsican Pine stand
Alongside the SALCA scanning, which will take priority this week, several experiments and field measurements have been discussed. Amongst these, an examination of the variability of hemispherical photography as a method for gap fraction analysis is started today and involves photos taken every day at plot 1 to cover a range of weather conditions. These images will then be processed by several different people to assess how much results vary.
Fisheye photo from plot 1
SALCA data from plot 1: 1550nm

We manage to get three scans in today at three different pine stands before retiring to our accommodation for the evening.

Oliver Gunawan
4th July 2011

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.

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

Wednesday, 25 May 2011

Trip 7

The weather is much cooler and overcast as the typical British spring is back. Luckily, we've managed to get a whole day without rainfall!

Welcome to the 21st century: there is no escape from your emails

Phenological Observations - Plot 2: The canopy here has not changed significantly since the last trip, although the undergrowth and in particular the brambles and bracken are still developing. Below is a graph displaying the differences in gap fractions in plot 2 over a month, using the fisheye photography. The light green line from the 19th April shows a larger gap than the dark green line from the 19th May, displaying how the canopy has filled in at different zenith angles (where 0 degrees is closest to vertical).

Canopy Gap Fraction Analysis from Plot 2

Phenological Observations - Plot 3: The canopy is still developing, however there is still very little undergrowth.

Clouds are starting to gather here, so we have decided to do a smaller resolution scan of the final plot, which will also produce a useful dataset for testing some of the code we have produced.

Chris with SALCA, surrounded by bracken at plot 4

25th May 2011
Oliver Gunawan

Thursday, 19 May 2011

Trip 6: Back on Track!

The 2 weeks of miserable weather has finally cleared and allowed us another day for field research. In addition, we have a new generator and back-up supply which gives us another level of security just in case.

Phenological Observations - Plot 2
: The Oak and Birch both appear to have reached full leaf and the canopy is now very well developed, with the whole of plot 2 looking significantly greener than 2 weeks ago. The brown/grey leaf litter is also gradually being replaced by green bracken and bramble bushes, some of which are producing flower buds.

Fisheye Photo of Plot 2 canopy
Phenological Observations - Plot 3: While plot 2 has a much more mature canopy; at plot 3 the Sweet Chestnut crowns are still developing, with significant growth even in the higher branches, but the canopy cover is not yet complete and leaves have not yet reached their full size. There is still very little undergrowth, but small shoots, grasses and moss are appearing near to the plot.

Oliver and Chris setting up SALCA at plot 3.
Phenological observations - Plot 4: Whilst the tree canopies have not changed much here, the undergrowth has exploded and ferns are now growing to head height all around the plot.
Fisheye plot from Plot 4 using the 1040 laser

19th May 2011
Oliver Gunawan

Wednesday, 4 May 2011

Trip 5: Problems with power

Another nice day for fieldwork, but the forecast shows that this could be the last chance in a while before rain sets in and it looks like the hottest April on record is starting to turn into much more typical UK Springtime weather. We arrive at plot 2 in good time, but have been starting to experience issues with our generator due to a loose connection in the setup. The generator and scan need to be restarted, but we do manage to complete a full recording.

Developing canopy at Plot 2
Phenological Observations - Plot 2: The canopy is developing in both Oak and Birch trees, with Oak leaves growing up to about 15cm in length. The undergrowth is greening up with most bracken plants, fully unfurled and some of the brambles starting to produce buds.

Oak leaf growth, Plot 2
At plot 3, the generator refuses to work and we have to admit defeat. After consultation with the manufacturers, it would appear that the invertor has blown and the whole generator must be replaced. As the unit is still under warranty, a new generator is on it's way.

Canopy at plot 3

Phenological Observations: Plot 3: First leaves are still not fully developed on higher branches of the Sweet Chestnuts, however the lower canopy is now very developed although there is very little undergrowth.

This has been a frustrating expedition, but we have resolved our power issues and have purchased a back-up power supply, which should help protect the circuit.

Oliver Gunawan
4th May 2011

Wednesday, 27 April 2011

Trip 4: The first Multi-temporal scanning of conifers using SALCA

The weather on our fourth trip is distinctly more overcast than our previous visits and the air is much cooler. Upon arriving at plot 2, the canopy has changed dramatically in just 1 week:

Phenological Observations - Plot 2:  There is leaf growth on the Birch and Rowan, but this is nothing compared to the Oak trees, which have completed first leaf and some of the trees are almost at full leaf in just one week. The canopy is now significantly covered with foliage. The undergrowth is greening up significantly, with large bracken unfurling and tangles of brambles appearing from autumn's leaf litter.

Fisheye photo at plot 2

The weather is improving as we make our way to site 3 and the sun is rapidly burning away all of the cloud cover.

Phenological Observations - Plot 3: The taller branches of the Sweet Chestnuts are starting to exhibit budburst with most even showing first leaf. The lower canopy consisting of younger trees is significantly more developed with leaves upto 50 mm in length. There is still little undergrowth, with just a few patches of moss appearing from the leaf litter

First leaf at plot 3

Phenological Observations - Plot 4: As expected, the coniferous canopy appears to have changed very little in one week, but the undergrowth has exploded into life. Bracken is the most common species here, growing in tall stalks in the shade, with bushier plants growing where there appears to be more sunlight. The maximum height is 1.3 m.

Setting up times are reducing with increased familiarity and we are currently taking steps towards analysis of the hemispherical fisheye photographs, which will provide a useful comparative result for Gap Fraction Analysis.

I am still very much on the menu as far as the mosquitos are concerned.

Oliver Gunawan
27th April 2011

Tuesday, 19 April 2011

The First multi-spectral scan of an evergreen stand

Delamere Forest, Cheshire, Plots 2, 3 and 4

Another bright and sunny morning has greeted our third field trip and as we are planning to visit all three sites today, our first scan in plot 2 is setup earlier than normal. We are also joined today by Jerry Joy, a Masters student at the University of Salford, who is conducting some research on how differing canopy cover affects GPS signals.

Jerry Joy with differential GPS receiver

Phenological observations - Plot 2: The oak trees are displaying a distinct difference from last week and are well into the first leaf stage of growth. The leaves on the birch trees are also continuing to grow and there is a definite 'greening' of the undergrowth, with a mass of brambles and bracken which are up to about 70cm.

First leaf on lower branch of Oak tree at plot 2

Phenological Observations - Plot 3: There is little phenological change in the larger Sweet Chestnut trees, with a small amount of bud burst visible. On the other hand, smaller Sweet Chestnuts and Rowans approx 3m high are exhibiting a first leaf and are already forming a secondary canopy. There is still very little undergrowth at this plot.

Fisheye photo from plot 4

Plot 4: The final plot is a mixture of Scots Pine and Corsican Pine. This is a predominantly evergreen stand, therefore phenological changes are not obvious in the canopy as new and old needles appear at the same time. This plot is to be used to compare against the deciduous plots and also to test how well SALCA can capture conifer trees in 3 dimensions.

Intensity image from plot 4 using the 1550nm laser

Phenological Observations - Plot 4: The ground is well covered by dead bracken and needle fall and there is a large covering of moss on the ground as well as vigorous bracken growth with immature plants growing up to 80m already.

The future of terrestrial laser scanning, today!

N.B. Mosquito repellent to be added to field equipment list.

Oliver Gunawan
19th April 2011