Comparison of output results from two programmes for hemispherical image analysis: Gap Light Analyser and WinScanopy
2010
Jarcuska, B.,Slovenska Akademia Vied, Zvolen (Slovak Republic). Ustav Ekologie Lesa | Kucbel, S.,Technicka Univ., Zvolen (Slovak Republic). Katedra Pestovania Lesa | Jaloviar, P.,Technicka Univ., Zvolen (Slovak Republic). Katedra Pestovania Lesa
We compare the results of the analysis of hemispherical images (of a broadleaved and a coniferous forest) obtained using the Gap Light Analyser (GLA) software and the results obtained by analyzing the same images with the aid of WinScanopy. The two packages were used to calculate relative total, relative diffuse and relative direct transmittance, canopy openness, and leaf area index. Our aim was to find out whether it is possible to compare the studies using different software packages for determining light conditions. The binary pixel classification of images of canopy and sky was performed automatically (in the case of Gap Light Analyser, using the SideLook programme). The threshold values determined by the SideLook programme were lower compared to the WinScanopy, which was also reflected in the evaluated output results. There was a strong positive correlation between the results obtained with the two software packages (R2 ranges from 0.814 to 0.999). However, when the Gap Light Analyser analysis was applied to the threshold values obtained with the SideLook, the output results mostly manifested systematic differences in comparison with the output results obtained using the WinScanopy. Using the same threshold value in both programmes, the differences between the output values were quite small (a minimum of 0.038 square m2/square m for LAI in the spruce forest and a maximum of 0.738% for total relative transmittance also in the spruce forest). The differences in some characteristics were statistically significant, on the other hand, both the photo series had only the identical direct transmittance values. The observed differences can be explained by differences in the calibration of the used camera-lens pair, different image registration techniques and different theoretical background and models used in the two software packages. It follows that it is also necessary to be aware of possible differences when comparing the outputs of the two compared software packages analyzing photos obtained applying the same methodical approach.
显示更多 [+] 显示较少 [-]