Canopy cover is measured using a variety of methods including spherical densiometers, funnels, moose horns, aerial photographs, and hemispherical images. As canopy cover requirements are tied to management activities, the accuracy and consistency of these measurements becomes increasingly important.

The goal of this project is to determine an efficient, reliable method to determine canopy cover from the ground. Additionally, it would be extremely advantageous to be able to predict future canopy cover, given the proposed removal of various trees within a plot.

Solution

After contacting several software firms and refining initial prototypes, we are currently testing a hemispherical image analysis package that operates on a Personal Digital Assistant (PDA). Thanks to Idruna Software Inc hemispherical images can be analyzed in Pocket Phojo. Idruna Software Inc has worked with us to develop the canopy filter within Pocket Phojo. This software package is commonly used by journalists who are interested in full image editing capabilities while on the go. This includes the ability to transmit the image wirelessly.

Images can be collected and analyzed using a digital camera and then transferring the image wirelessly or by physically moving the memory storage card from the camera to the PDA. Alternatively, images can be collected using a camera in the PDA. The Pocket Phojo software can then be used to measure percent canopy closure of the existing plot. The software also allows the image to be edited, where the user can "cut" the trees of their choice and recalculate the canopy cover.

Hemispherical images are among the most accurate means to measure canopy closure. Additionally, image editing appears to be the only means to rapidly predict after harvest canopy cover. While canopy cover can be estimated using basal area, these estimates are commonly inaccurate. An additional benefit includes the potential to accurately estimate forest biomass. The Fire Research Lab found a high correlation between hemispherical image analysis and kg/m^2 of woody biomass. This correlation will be verified through a cooperative effort between SDTDC and Oregon State University.

The following images are screen captures from a PDA using Pocket Phojo. The first image is a hemispherical view of a forest canopy from below. The next image is the conversion of the original into black and white pixels. The amount converted to black pixels is dependent upon the threshold level, which can be adjusted. The software analyzes the image and determines the canopy cover to be 76%. The final picture demonstrated the ability to "cut trees" from the image to estimate the canopy cover after the proposed stand treatment. After the estimated treatment the canopy cover is 63%.

 

In addition, the zenith angle can also be adjusted, where 0 degrees is the center of the image, directly overhead, and 90 degrees is the edge of the image. In this case the zenith angle is set at 45 degrees. You may have noticed that it is possible to erase the canopies of tress that are not expected to be removed. This issue can be reduced by lowering the zenith angle.

The image, the analyzed image, and the relevant data (Threshold, Zenith, Location, Percent Cover) are all easily saved with this system. This allows new analysis or verification without returning to the site.

Sources

The software can be ordered online at the Idruna Software Inc web site. A variety of camera and lens options are available, including those by Canon, Nikon, and Olympus. Digital SLR's are the most likely option, however, there are a few point and shot cameras that will accept a 180 degree fisheye lens. Costs will vary, but the software, camera, and lens will run around $1,700 depending upon the equipment purchased. If you are interested using this method and have questions regarding what equipment you can use, contact us and we'll be happy to help.

The following information was compiled to assist in purchasing a lens/camera combination. This information was updated in December, 2013.

What modern lens and camera does Idruna Software recommend?

In the year 2007 the price of a hemispherical lens ran at a high price which has now been reduced. Today's models cost less than three hundred dollars. Idruna Software recommends a lens made by Rokinon. The Rokinon 8mm F/3.5 and F/2.8 models were developed in 2012 and have made the fisheye lens cheaper. These lenses can be ordered or purchased from any electronics store. The Rokinon 8mm lens can attach to any SLR type camera. The manufacturer of the lens recommends the Pentax Digital SLR K series cameras with the Rokinon 8mm lenses.

There are plenty of cheaper (less than $200) ‘fisheye’ lenses, why can’t those be used?

There is a difference between a true hemispherical lens and a fisheye picture, a hemispherical lens produces a proportional image while a fisheye lens distorts an image. The result is a skewed image. A true hemispherical lens creates an image in which the distance along the radius of the circular image is proportional to the zenith angle. Cheaper lenses cannot make a proportional image. A true hemispherical lens has a small focal length and requires a large sensor to capture a full 180° image.

Why can't a point and shoot camera take a true hemispherical photograph?

Most point and shoot cameras have small sensors. There are engineering difficulties in delivering a small focal length combined with the very small sensor commonly found in typical point and shoot cameras. Most images taken with a point and shoot camera that appear hemispherical are modified with a computer program.

How do I identify a true hemispherical lens?

Hemispherical lenses have a focal length ranging from 8 to 12 mm. A lens with a focal length larger than this is not a true hemispherical lens or will not capture a full 180 degrees. Point and shoot camera lens focal lengths are a minimum of 24mm. Any lens attached to the top of this lens increases the focal length. The small sensors of the point and shoot cameras require larger focal lengths to let more light in. A large sensor and lens needs a smaller focal length. SLR cameras can capture images in low light by having a large sensor.

This project was initiated by the proposer, Mike Malone, of the Mount Hood National Forest in Region 6.