Hemisfer ist ein Programm zur Schätzung vom Blattflächen-Index (BFI, oder LAI in Englisch: leaf area index) und vom Licht-Haushalt aus hemispherischen Photos.
See the full version of these pages in English.
Ursprünglich wurde Hemisfer entwickelt, um verschiedene Berechnungsmethoden fürs BFI zu vergleichen. Wir wollten zuerst einzelne freie oder kommerzielle Programme dazu verwenden, aber sie gaben gewisse Unterschiede in den Resultaten, die wir gar nicht erklären konnten. Deshalb haben wir 4 verschiedene Methoden in einem einzigen Programm gepackt. Beim Versuch, diese Methoden zu optimieren, sind wir auf noch eine weitere Methode gekommen. Schritt für Schritt sind hat sich so ein Produkt ergeben, das wirklich interessante, zum Teil weltweit neue Eigenschaften besitzt.
Weil die Entwicklung dieses Programmes an einem öffentlichen Forschungsinstitut gemacht wurde, ist es unser Interesse, unsere Resultate zugänglich und bekannt zu machen.
Hemisfer steht frei zum Herunterladen, in einer Version, die voll funktionsfähig ist. Die einzige Begrenzung ist, dass damit Bilder nur bis einem Durchmesser von 1000 Pixels damit zu analysieren sind. Dies reicht völlig für Versuche und Übungen, und die erreichte Genauigkeit sollte in vielen Fällen ausreichend sein. BFI Schätzungen werden ja relativ wenig von der Bild-Auflösung beeinflusst. Viele andere Faktoren spielen eine grössere Rolle, und da bietet Hemisfer wertvolle Lösungen und Anweisungen (dank einer integrierten Hilfe, die aber erst auf Englisch erhältlich ist).
Die Begrenzung in der Bildgrösse wird aufgehoben, sobald das Programm registriert wird. Für nicht-kommerzielle Anwender ist die Registrierung billig (80 schweizer Franken). Für kommerzielle Anwendungen ist sie teurer, aber noch günstig (500 schweizer Franken). Registrieren und zahlen ist einfach und wird voll elektronisch abgewickelt. Registrierte Kunden dürfen unsere technische Unterstützung in Anspruch nehmen (siehe unsere Gebrauchs-Bedingungen). Eine FAQ Seite steht auch zu Verfügung.
Um Hemisfer zu testen haben wir zudem ein Programm namens Canopix entwickelt, das künstliche Bestands-Bilder generiert. Die Bilder werden zufällig auf der Basis von eingegebenen Parametern produziert. Diese wählebare Parameter sind: BFI, durchschnittlicher Blattwinkel und Steigung des Bodens. Canopix steht ganz frei zum Herunterladen bereit.
Hemisfer is a software designed to estimate the leaf area index (LAI) and the light regime from hemispherical (or wide-angle) photographs. It includes the following capabilities:
- Runs under Windows (XP, Vista, 7 or 8).
- Hemisfer is multilingual. Following language files are available at present: English, French, German, Italian, Spanish, Swedish, Russian, Ukranian, Turkish, Croatian and Romanic.
- Hemisfer has an extensive online help (in English only).
- It works with JPEG, TIFF, PNG, BMP or GIF picure files with 2, 8 or 24 bits per pixel.
- Pictures can be analysed either individually (interactively) or as a batch (user-unattendend).
- Information about these pictures can be stored in plain-text (understandable) parameter files and retrieved from there.
- It can also import HVS parameter files generated by HemiView, a software by Delta-T.
- Several methods are available to estimate the LAI from the light transmission in rings around the zenith:
- Miller (1967)
- Miller (1967) as implemented in the Li-Cor LAI-2000
- Lang (1987)
- Norman & Campbell (1989)
- Thimonier et al. (2010)
- There is the possibility to take the effect of the ground slope into account by using the algorithm of Schleppi et al. (2007).
- An algorithm (based on Chen & Cihlar, 1995) is included to separate the gaps out of clumped canopies and to correct for their effect on the LAI calculation. The method of Lang & Xiang (1986) is also available to correct these clumping effects, as well as the method of Walter & Torquebiau (2000), which corrects for both ground slope and canopy clumping.
- It is possible to calculate the light regime at the location where the photograph was taken. The incident light (diffuse and direct) is modelled at a minute time step over the year and combined with the picture as a mask to produce a below-canopy light regime. Results can be diplayed at different time steps within the daily and yearly cycles.
- The software is able to calculate the threshold to distinguish between white (sky) and black (canopy) pixels according to the methods of Nobis & Hunziker (2005) or of Ridler & Calvard (1978). The γ (gamma) value of the picture is thereby taken into account, which gives objective and reliable results.
- Results provided in standard file formats, easy to import into softwares used for further processing (spreadsheets, statistical softwares, graphical softwares).
Note: the free downloadable software is limited to an image analysis within a radius of 500 pixels. This limitation is removed upon registration.
- If you want to know what's new in the latest version, you can check the release notes (see tab Release notes).
- Download the current version (2.2) as Windows installation package. If you are nor sure to have administrator rights on the target computer, you may prefer to download it as a ZIP file for manual installation.
- Installing Hemisfer is then very easy: double-click on the downloaded intallation file and follow the instructions that appear. A shortcut will be placed in the start menu and on the desktop. If you downloaded instead the ZIP file, it is not complicated either: uncompress all files into a folder of your choice. You can then make a shortcut of Hemisfer.exe in your start menu or on your desktop.
- Hemisfer is provided with a few language files. You can therefore choose another language at any time without having to reinstall anything, not even to leave the software.
- After installation, the free downloadable software is limited to an image analysis within a radius of 500 pixels. This limitation is removed upon registration (see tab Registration and payment).
Canopix is a companion software to Hemisfer for generating random canopy pictures. Its is available as freeware without warranty or support.
Canopix (version 2.0) can be downloaded as executable file and does not need any specific installation. It runs on Windows Vista or more recent.
Hemisfer is shareware. It can be redistributed freely as long as no modifications are done to the original distribution package (except for adding lens parameters or adding new language files). The author retains the copyright of the software.
Hemisfer was developed with great care, but it is provided with absolutely no warranty. After installation, the free downloadable software is limited to an image analysis within a radius of 500 pixels. This limitation is removed upon registration (see tab Registration and payment).
Two different registrations are available, either commercial or non-profit. A non-profit registration must be justified by indicating in which context the software will be used. If you plan to use Hemisfer for teaching, note that a commercial registration is also valid as a multi-user educational registration.
Users (registered or not) must include a proper citation if they write scientific articles including results obtained using Hemisfer. Because the status of our own publication will change with time, the proper citation must be checked when the user's article is submitted or resubmitted after revision. Currently valid citation is:
- Schleppi P., Conedera M., Sedivy I., Thimonier A., 2007: Correcting non-linearity and slope effects in the estimation of the leaf area index of forests from hemispherical photographs. Agric. Forest Meteorol. 144: 236–242.
- Thimonier A., Sedivy I., Schleppi P., 2010: Estimating leaf area index in different types of mature forest stands in Switzerland: a comparison of methods. Eur. J. For. Res. 129: 543-562
An help file is provided with Hemisfer. If they don't find answers to their question in this help, registered users can ask for e-mail support. There is however no warranty about response time and users may only be referred to a page of frequently asked questions (see tab Frequently asked questions). General questions about hemispherical photography and LAI estimation may not be answered if they are not directly related to the use of Hemisfer.
Yet another LAI software? Why?
See our motivations on our home page!
Why this name Hemisfer?
The software analyses hesmispherical photographs, but the author decided to change ph to f because so the name includes the letters ISF, which is the French abbreviation for LAI (Indice de Surface Foliaire).
On which operating system can I use Hemisfer?
Hemisfer runs on Windows, either Vista, 7, 8 or 10. It may still work on XP, but without warranty. It does not (yet) with Wine under Linux.
I try to install Hemisfer but get a warning that the source of the software is not certified. Should I care?
A certificate authentifying a software is expansive and needs to be renewed regularly. Having such a certificate for Hemisfer would mean roughly doubling its price. I prefer to keep the price as it is and to continue offering a free unregistered "exercice" version.
Is it possible to use this software for pictures taken with non-fisheye lenses (e.g. wide-angle 16mm, or even 28mm)?
Yes, the definition of the lens geometry includes this option. If the field of view is less than about 45°, however, the mean angle of the foliage becomes hard to estimate from the pictures. In this case, it may be better to restrict the foliar angle to a narrow plausible range in order to get a good estimate of the LAI.
Can I calculate the light regime prevailing at the point where the photograph was taken?
Yes. Hemisfer is able to calculate the transmission of direct and diffuse radiation through the canopy. For the incident (above-canopy) radiation, it uses a model requiring only 2 parameters: the average direct and diffuse radiation of the atmosphere.
I can't use the help: it tells only "Navigation to the webpage was cancelled". What can I do?
You probably installed Hemisfer and its help file on a network server. This is a known restriction of Windows' help files in general, see <http://support.microsoft.com/kb/896054>. Hemisfer tries to find out if such a restriction applies and then shows the help from the web <http://www.schleppi.ch/hemisfer/help>. This may however not work in every situation. If you get the above-mentioned error message, try installing Hemisfer on a local drive. There is no problem with the registration if you change the location of the programme's folder. Or you may prefer to rename the help file: if it cannot be found, then the online help will be shown instead.
I have an error message about lenses, that a file cannot be read. Why?
A lens name is defined in a parameter file (.INI), but the lens parameters are taken from the Lenses.csv file, which must be either in the same folder as the Hemisfer software or in the same folder as your parameter and picture files. A lens file with some of the most commonly used lenses is included in the download package.
I changed my lens file and saved it again in the same folder, but then Hemisfer no longer finds any lens parameters. What's wrong?
The CSV extension of Lenses.csv means comma-separated values. However, if you edit such a file with a spreadsheet, it may happen that it uses different separators like semi-colons. Hemisfer tries its best to cope with such differences in formats, but does not always succed. Check the format of your lens file in a simple text editor (and compare it with the original lens file inluded in the download package).
After saving a file, there is an error message about an invalid function call, but actually the file is saved correctly. What's the problem?
This may be due to a shortcut (in the start menu or on the desktop) that indicates a wrong starting folder. Try modifying this shortcut (right-click, then "properties"), or replace it by a new one (drag Hemisfer.exe and drop it while holding the CTRL and SHIFT keys down).
The programme always asks me if I want to save parameters. Can I get rid of these messages?
In the options about results, you can define if your parameters are saved always, after confirmation or only by request from the user. Chose the option you prefer.
I would like to calculate only the threshold for a whole batch of pictures. What's the best way to do this?
Open all the files, remove all LAI methods from the options, and set the 'saving parameters' option to always. Start then the analysis and the results will be saved both individually in the INI files and together in the results file.
I have problems to close Hemisfer when the results are still displayed in my text editor. What's wrong?
This problem is known for some programmes which do not handle properly the communication when they are called. If you are able to do it: modifiy the association of the file type (e.g. *.DAT) so as to avoid DDE calls. If this is too complicated for you (because you are just a normal human being): close the results window first and... sorry about it (but don't blame me for bugs of other softwares).
I change the γ (gamma) value but this has no effect on the LAI calculations. Why?
The γ value is used only to calculate the threshold. This means that it affects the LAI only if you let Hemisfer calculate the threshold. If your threshold is fixed or set via a parameter file, then γ has no effect.
Can I open Hemisfer several times to compare pictures?
Yes, it works, allowing you to compare also differents parameters for the same picture.
I set rings of 15° width, but in the results the angles are 10°, 23.3° etc. Why these numbers?
These are weighted means of all angles within a ring.
Picture analyses from a same site don't appear in the same summary in the results. Why?
The files are treated in alphabetical order. All files from a site should thus appear together in the alphabetical list and then the summary will be based on the site name as given in the parameter files. Just give file names accordingly, for example by putting the site name or abbreviation at the beginning.
The correction of slope and non-linearity (Schleppi et al., 2007) increases the calculated LAI but it often increases also the proportion of gaps in vertical projection (Fmv). Isn't this contradictory?At first glance yes, because more LAI, i.e. more foliage should leave less gaps open. However, the correction of slope and non-linearity often gives also a higher leaf angle, which lets the leaves appear smaller when looking up vertically.
2 May 2017: Hemisfer 2.2
A new method based on a robust regression (Gonsamo et al.) has been added to Hemisfer for the calculation of the leaf area index. In its results, version 2.2 also includes the canopy openness as the solid angle proportion of visible sky. (So far, only the proportion of sky in azimuthal equidistant projection was reported as "transmission".) Further, a small inaccuracy in the calculation of the diffuse light regime on a slope was corrected. There was a relative error in the diffuse light index (DLI) of 0.04% by % of slope. This difference should thus be taken into account when comparing older results with results obtained with this new version. When the ground is flat (slope = 0), there is no change in the obtained results.
2 May 2017: Hemisfer 2.17
An error appeared on some computers running Windows 10, preventing Hemisfer to start. Version 2.17 corrects this. It adds also a new language to the user interface: Portuguese.
7 April 2016: Hemisfer 2.16
Since version 2, changing the gamma value during a working session did not have any effect on the threshold calculation according to Nobis & Hunziker (as it should have). In a normal use, gamma is not changed after it has been defined, and in such cases the thereshold was always correct. Version 2.16 now takes any change in the gamma value into account, also within a working session. Further, the p-value for the clumping calculation after Chen & Cihlar is now more strictly defined: in earlier versions, it would have been possible to have a (meaningless) value above 1.
22 March 2016: Hemisfer 2.15
The behaviour or the colour frequency graph (histogramm) was improved for Hemisfer version 2.15, The colour to be ignored is now really ignored also in the theresholding after Ridler & Calvard (1978). Further, a bug in the display introduced in version 2.14 (missing or misplaced rings etc.) was corrected.
15 March 2016: Hemisfer 2.14
Hemisfer version 2.14 brings minor improvements to the user interface.
25 November 2015: Hemisfer 2.13
Hemisfer version 2.13 corrects a bug that produced an error message "index out of range" depending on the lens and on the number of rings. The handling of site coordinates had also to be corrected because "south" and "west" indicators got lost.
25 September 2015: Hemisfer 2.12
Hemisfer version 2.12 brings small improvements to the user interface and corrects a bug that generated an error message when trying to calculate the light regime for sites a high latitudes (>59°).
4 August 2015: Hemisfer 2.11
Hemisfer version 2.11 corrects two small bugs that appeared in some cases when the number of rings was reduced. One of these bugs generated an error message during threshold calcualtion, the other one put the software into a non-responding state (endless loop) when using a single ring but several sectors.
4 May 2015: Hemisfer 2.1
So far, the slope correction used in Hemisfer (based on Schleppi et al., 2007) produced LAI values expressed by unit of inclined ground. Starting with version 2.1, Hemisfer now gives the slope-corrected LAI per unit of projected (horizontal) surface, which is the general standard. Results from older versions can be converted to this improved calculation method by dividing them by the cosine of the slope. Relative comparisons with the same slope and all calculations of light regimes are unaffected by this change.
2 March 2015: Hemisfer 2.07
Hemisfer 2.07 corrects an error that occured when trying to apply a per-ring threshold on a picture that was actually already black and white only.
10 February 2015: Hemisfer 2.06
When trying to save a new lens geometry, the software appeared to freeze (and came back to life only after switching to another application and back). Hemisfer 2.06 corrects this problem.
10 January 2015: Hemisfer 2.05
Hemisfer 2.05 corrects a bug that the window of the software could have been placed outside the screen (which resulted in an icon in the task bar but no visible Hemisfer window on the screen).
3 November 2014: Hemisfer 2.04
In some cases, Hemisfer caused errors because it did not get the permission from Windows to write the sites and lenses files into its own folder. These files are thus now located within the user's AppData folder (except if the programme folder is writeable for everyone, e.g. in a portable installation on a removeable drive). Version 2.04 further corrects an error message that could appear when sectors were used at the same time as the calculation of the light regime.
2 October 2014: Hemisfer 2.03
Version 2.03 corrects a problem with the colour to be ignored. When painting zones to be ignored on a picture, it is possible to specify a tolerance because non-conservative picture formats like JPEG do not keep the colours exactly. The calculation of this tolerance was wrong (too much tolerance) in the preceding versions 2.x.
4 September 2014: Hemisfer 2.02
Depending on the country and language settings of Windows, text files (lenses, sites, parameters) were not always correctly read. This is corrected by the new release. Automatic saving of parameter (INI) files did not work properly and this is now also improved. Further, it is now possible to install the new version without uninstalling the old one (existing lenses and sites files are thereby preserved).
21 August 2014: Hemisfer 2.01
This corrects a a bug in version 2: diffuse normal radiation was not properly reset between the analysis of pictures within a batch analysis.
18 August 2014: Hemisfer 2
Version 2 is finally available. It has two main new features. First it can now make a local analysis of the light regime at the point where a hemispherical picture was taken. Second, it runs properly on newer versions of Windows (7, 8). Older registrations are still valid for this version. Note that INI parameter files from version 2 will not be usable with older versions of the software. The compatibility is only "upward": older parameter files can be read in Hemisfer 2.
24 February 2011: Hemisfer 1.53
This version corrects a bug that affected the calculation of the lens geometry for some lenses. Nikon FC-E9 was slightly affected, Raynox DCR-CF185 more. Users who worked with these lenses or with parameters not provided with Hemisfer are advised to check the results obtained with previous versions as they may be inaccurate.
16 August 2010: Hemisfer 1.52
This version corrects the fact that the suffix .INI got lost when parameter files were generated automatically. It also avoids an unnecessary error message when the foliage angle is estimated to 90°.
29 June 2010: Hemisfer 1.51
There was a bug in the previous versions, affecting the correction of canopy clumping after Chen & Cihlar (1995) when there was a colour to be ignored in the photographs. These two options worked well separately, but when combined they gave too high LAI values. Version 1.51 corrects this bug.
27 April 2010: Hemisfer 1.5
With this version, the gap fraction in vertical direction can now be calculated out of the obtained LAI and leaf angle. This gives also the canopy cover, which is defined as the proportion of the forest floor covered by the vertical projection of the canopy. The same calculation can be done for the fraction of large gaps. Further, it is now possible to constrain the mean leaf angle to a defined range. The user interface received small improvements.
30 July 2009: Hemisfer 1.44
This version adds the possibility to get (uncorrected) LAI values per sector.
4 June 2008: Hemisfer 1.43
This version corrects two small bugs (which affected the user interface, not the calculations). Results without slope correction are now always displayed, also (for comparison) along with slope-corrected results.
15 April 2008: Hemisfer 1.42
This version corrects a bug in the error handling that caused the software to crash when encountering some error situations (e.g. image file not found).
27 August 2007: Hemisfer 1.41
This version corrects a problem with the synthesis of results across several sites. The data were not correctly reset between sites and the systheses of subsequent sites were therefore wrong. The results of individual pictures as well as the synthesis of the first (or single) site analysed in a run were right and thus remain unchanged in this version.
6 June 2007: Hemisfer 1.4
Two new methods were integrated for correcting the effect of canopy clumping (Lang & Xiang, 1986) or clumping and ground slope (Walter & Torquebiau, 2000) . These methods can now be chosen instead of the already available corrections of ground slope (Schleppi et al., 2007) and clumping (Chen & Cihlar, 1995). This latter method was improved and may thus give slightly different results compared to previous versions. The clumping factor Ω (Omega) was used correctly, but given as its inverse 1/Ω in the results; this is now corrected. The lens formula in the help was corrected (radius r and angle Θ were exchanged).
The internalisation was improved and Hemisfer should now run without problem on computers using the comma instead of the dot as decimal separator. Parameter files may however not be exchanged between one system and the other. Further minor improvements were made to the user interface.
9 May 2006: Hemisfer 1.3
Tests with artificial pictures revealed a small error in the weight of the different rings when zones of the picture have to be ignored. This error affected pictures with zones painted with a color to be ignored as well as pictures where the soil is visible in the analysis rings because of a ground slope. Users who analysed such pictures are advised to repeat the calculation with the version 1.3 and see if this error is important in their case. If no pixels were ignored, then the results are not affected. Version 1.3 also corrects a bug causing the programme to crash when minimised. It brings further minor improvements to the error handling and to the user interface.
28 October 2005: Hemisfer 1.2
Version 1.2 now includes a graph of colour frequencies, the definition of the north on the picture, the support of lenses with a viewing angle less than 180 degrees and a second thresholding method (Ridler & Calvard, 1978), beneath the already provided method (Nobis & Hunziker, 2005). It also corrects problems encountered when changing the language and brings minor improvements of the user interface, including a better handling of run-time errors.
6 September 2005: Hemisfer 1.11
The parameter save button was not always enabled as it should have been. Revision 1.11 corrects this.
30 August 2005: Hemisfer 1.1
Version 1.1 mainly corrects a bug which caused a wrong display of the picture when zooming out, or even let the programme crash. Some details of the user interface were also improved.
11 July 2005: Hemisfer 1.0
Version 1 of Hemisfer is released as a shareware for the calculation of LAI from hemispherical photographs of plant canopies.