THE INFLUENCES OF IMAGE CLASSIFICATION BY FUSION OF SPATIALLY ORIENTED IMAGES

Image classification of either supervised or unsupervised approach is an essential tool to categorise the unidentified pixel in an image to the thematic or the spectral separable classes in the remote sensing science. The informational utility of single-image classification in somehow is limited by either the spatial or spectral resolution, due to the physical trade-off between the resolutions of imaging system. In order to integrate both high spatial and spectral resolution in a single image, the technique of image fusion may be employed. This paper investigates the influences of the multispectral image which is fused with the spatial-oriented image, on the thematic accuracy and the resultant clusters of supervised and unsupervised classification respectively. Through two examples of spatial-oriented images: SPOT panchromatic and scanned aerial images, two respective SPOT multispectral images were fused by intensity-hue-saturation (IHS), principal component analysis (PCA) and high pass filter (HPF) fusion methods. All the images were then classified under the supervised classification approaches of maximum likelihood classifier (MLC) and the unsupervised approaches of ISODATA clustering. Using the classified result of the parent (original multispectral) image as a benchmark, the integrative analysis of the overall accuracy and the numbers of clusters indicated a certain degree of improvement in the classification from using the fused images. The effect of various resolutions for image fusion is also presented. The validity and limitations of image fusion for image classification are finally drawn.