本論文針對IKONOS衛星影像的特性來進行影像融合,主要目的是將ㄧ張高解析度的PAN影像(high-resolution panchromatic)和四張低解析度的MS影像(low-resolution multispectral),融合成一張兼具PAN影像中的空間資訊優勢和MS影像中的光譜資訊優勢的融合影像。 本研究提出一個修正模型,利用調整BT(Brovey Transform sharpening method)與HIS(intensity-hue-saturation transform method)之間的融合參數值,並以ANN(Artificial Neural Network)方法來協助達成融合參數的調整最佳化,來達到影像融合,同時具有精確的空間資訊及色彩資訊。
The IKONOS satellite produce a PAN image (high-resolution panchromatic) and 4 MS images (low-resolution multispectral) simultaneously. In this thesis, an image fusion method is proposed to fuse the PAN and MS images. In the proposed method, not only the edge information in the PAN image but the spectral information in the MS images can be preserved. In the proposed fusion method, ANN (Artificial Neural Network) is used to optimize the fusion parameters which involve the parameter adjustment with BT (Brovey Transform sharpening method) and HIS (intensity-hue-saturation transform method) HIS (intensity-hue-saturation transform method). Based on the simulation results obtained in this study, the proposed fusion method indeed improves the perceptual quality in both edge and spectral information