A spot of good news for plant ecologists

Thursday, December 19th, 2013
Black Spot

Pirates shuddered at the thought of black spots - they signified your imminent bloody demotion from pirate hierarchy. But for ecologists, the recently released Black Spot Leaf Area Calculator is a cause for celebration. Developed by NCBS doctoral students Varun Varma and Anand M. Osuri, Black Spot quantifies the surface area of an individual leaf. Leaf area measurements are used to derive a number of metrics used extensively in ecological studies.

There's more to a leaf

A leaf can furnish a lot more information than meets the eye. While a leaf's vein network and its patterns of arrangement on a stalk often help ecologists identify plant species, other morphological and physiological leaf characteristics give insights into how these plants respond to and interact with their surroundings. Traits like leaf toughness for example, are a sign of plant defences against pests. Others like leaf nutrient levels are an index of photosynthetic activity. Similarly, ecologists often measure the surface area of a leaf, since it is an indicator of the environmental stresses that plants face and, at a deeper level, of the ecological and evolutionary strategies they adopt to overcome such hardships. Various other important metrics in plant ecology can also be derived from leaf area measurements.

So how does one measure leaf area? Specialized leaf area meters exist, but affordability is often a worry here. To cut costs, most scientists take digital images or scans of leaves and utilize image-processing software to obtain manual leaf area measurements from scans. In these cases however, users have to also invest more time to calibrate images by measuring surfaces of known area frequently. And because leaf area calculations require large samples sizes (leaf area across individual plants can be very variable), quantifying leaf area ends up being a tediously long-drawn procedure.

That is precisely why Varma and Osuri came up with Black Spot: a free image-processing software that produces automated leaf area measurements from scanned images of leaves. The software requires very minimal input from users, making it a blessing for large data sets.

How it works

Black Spot is based on the principle of band ratios, a theory widely used in remote sensing techniques. Every surface is different in the way it reflects the amount of light that falls on it - be it our bodies or leaves of trees. In the case of scanned leaf images, the leaf and white background surfaces reflect light differently. White has very similar (high) values for red (R), green (G) and blue (B) bandwidths, and the ratios between these colors (R:G, R:B and B:G) tend towards a value of 1. On the other hand, leaves have very different values of R, G and B (some low and some high) and the ratios diverge from 1.

What of leaf shadows created in the scanned images? "The shadows of leaves, while being of lower intensity, have band ratios which are very similar to paper," says Varma. Thus ratioing also effectively tackles the problems of dark shadows in scanned images. "The classification based on band ratios also means that the rulesets do not need to be modified for different types of leaves," the authors add. This eliminates the need to frequently calibrate images during scanning.

Written in Python, Black Spot first designates an image into leaf and non-leaf area using these differences in band ratios. The software then uses a smoothening filter to correct misclassified pixels within the leaf surface. Most importantly, the software extracts information from image meta-data in its EXIF file to detect image scale, thus eliminating the need to measure the leaf's scale manually.

And exactly how fast is it? Tests revealed that manual processing of a hundred leaf images took nearly 3 hours and thirty-six minutes, which Black Spot accomplished in exactly 20 minutes.

Varma and Osuri's colleague Yadugiri V.T. thinks Black Spot is "brilliant" for this very reason. "It drastically reduces the time needed to get leaf area. The output is a CSV file so there's no data entry required too."

While the speed is the highlight, Karthik Teegalapalli scores the application high on simplicity as well. "I have used it to estimate leaf area for several species from Arunachal Pradesh where taking complex equipment is often not an option," says Teegalapalli, Research Scholar with the Nature Conservation Foundation, Mysore and a member of Researchers for Wildlife Conservation, India.

"The thought and attention to detail that has gone into building the program, I think, is remarkable," says Yadugiri. It went through various stages including an initial one which necessitated placing the leaf against a background containing a small, black reference square while scanning - which lent the software its current name, Black Spot. Varma and Osuri began developing the script in mid-2011 and tested and fine-tuned the software till early 2013. "The main challenge was making the best of our limited programming skills to create an image processing software that others might actually choose over other wonderful software such as ImageJ or GIMP," say the authors. "It has been a rewarding experience, particularly because of the good results we started to get quite early on in terms of accuracy and speed of results across a wide variety of species and leaf types," they add.

258 downloads... and counting

And this success is apparent in the number of downloads so far: 258 Black Spot downloads across forty-four countries (as on 20th December 2013, see http://sourceforge.net/projects/blackspot/), almost half of which are from India. The authors also enrolled the software in the Bitplane Excellence Award competition. It cleared the first round of shortlists for the award, and has been one of the top-rated entries in the competition.

The only drawback that Yadugiri notices is that sometimes the software does not detect very small leaves on screen. "While BlackSpot works very well on large and medium-sized leaves, very small or very thin leaves are still a bit of a problem - like blades of grass," she says. For applicability in the field, she thinks it would be great to have another version of Black Spot which works on photographs taken with a camera, rather than on scanned images of leaves, she says. "But I think that's something Varma and Osuri are already working on."

They do plan to add more features to Black Spot to increase its applicability. "Our next goal is to create a graphic user interface for Black Spot," say the authors. "We also plan to expand the functional capability of Black Spot by including features to collect data on leaf shape, perimeter-area ratio and so on." With such additions, the Black Spot Leaf Area Calculator is sure to feature in every plant ecologist's list of coveted open-source software.

(Story edited by Geoff Hyde)

-----------------------------------------------

The Black Spot Leaf Area Calculator is hosted online on the NCBS website. The paper published in the journal Plant Ecology can be accessed at http://link.springer.com/article/10.1007%2Fs11258-013-0273-z.

For a free download of the author-generated paper, please visit Varma's NCBS page at http://www.ncbs.res.in/node/918.

Comments

Post new comment

The content of this field is kept private and will not be shown publicly.
CAPTCHA
This question is for testing whether you are a human visitor and to prevent automated spam submissions.
Image CAPTCHA
Enter the characters shown in the image.

Bookmark and Share