Vegetation Classification Accuracy Research Articles

Normalization of bi-directional effects in aerial CIR photographs to improve classification accuracy of boreal and subarctic vegetation for pollen-landscape calibration

Bi-directional effects were quantified and normalized in colour infrared (CIR) aerial images in order to produce high accuracy, high-resolution vegetation maps for the needs of pollen-landscape calibration. Three areas in northern Finland with different vegetation types were studied. The brightness variations of the most typical land cover type of each study site were quantified by examining the change of mean DN values of the reference sample areas on three separated wavelength bands in the overlapping images. Correlations between DN values and the scattering angles of reference sample areas were taken into account while normalizing bi-directional effects. The normalization was validated by classifying the images both before and after the normalization and assessing the classification accuracy. The accuracy increased on every study site after normalization. The highest increase was achieved in areas of uniform mountain birch forest. Increase was considerably lower on areas of fragmented vegetation. The resulting land cover classifications were used as landscape information in pollen-landscape calibration.

Integrated vegetation classification and mapping using remote sensing and GIS techniques

NOAA-AVHRR data have been more and more used by scientists because of its short temporal resolution, large scope, inexpensive cost and broad wave bands. On macro and middle scale of vegetation remote sensing, NOAA-AVHRR possesses an advantage when compared with other satellites. However, because NOAA-AVHRR also possesses problem of low resolution, data distortion and geometrical distortion, in the area of application of NOAA-AVHRR in large scale vegetation-mapping, the accuracy of vegetation classification should be improved. This paper discuss the feasibility of integrating the geographic data in GIS (Geographical Information System) and remotely sensed data in GIS. Under the environment of GIS, temperature, precipitation and elevation, which serve as main factors affecting vegetation growth, were processed by a mathematical model and qualified into geographic image under a certain grid system. The geographic image were overlaid to the NOAA-AVHRR data which had been compressed and processed. In order to evaluate the usefulness of geographic data for vegetation classification, the area under study was digitally classified by two groups of interpreter: the proposed methodology using maximum likelihood classification assisted by the geographic database and a conventional maximum likelihood classification only. Both result were compared using Kappa statistics. The indices to both the proposed and the conventional digital classification methodology were 0.668 (very good) and 0.563 (good), respectively. The geographic database rendered an improvement over the conventional digital classification. Furthermore, in this study, some problems related to multi-sources data integration are also discussed.

多時期のランドサットＴＭ画像を用いた湿原植生分類

Wetland monitoring and especially wetland vegetation classification are crucial for preserving valuable wetland ecosystems. The development of remote sensing technique for wetland monitoring is of urgent necessity. In order to improve the accuracy of vegetation classification, we have investigated the wetland vegetation classification using multi-temporal Landsat TM images. Because the growth pattern of a wetland vegetation changes according to the vegetation type, we can used this difference of temporal growth pattern which appear in the multitemporal images for classifying the vegetation types. In order to clarify this temporal growth pattern of wetland vegetation types, we have conducted sampling experiments to measure the biomass growth during the growing season. And also spectral reflectance measurements were conducted to see the spectral difference between the vegetation types as well. As the result of supervised classifications using the multitemporal Landsat TM image, an accurate wetland vegetation classification map has been produced.