保留邊緣濾波器評估指標之研究

Abstract

影像中的邊緣線在製圖應用與判釋上均極為重要，當帶有雜訊之影像進行平滑化處理以過濾雜訊時，並不希望因為邊緣線模糊而導致後續處理，如偵測邊緣線、特徵粹取等，產生困難；因此當帶有雜訊之影像以濾波器進行前處理時，邊緣線的保留程度即為待討論的重點之一。 許多濾波器因而被設計成同時具有過濾雜訊與保留邊緣線功能，如最鄰近勻化濾波器(Symmetric Nearest Neighbor Filter, SNNF)、Kuwahara濾波器、MCV濾波器等。 本文中以兩類實驗驗證評估指標之功能:1. SPOT多波段影像(PAN與XS1、XS2、XS3);2. Lenna人物像加入不同隨機雜訊後再經過八種濾波器(均值濾波器、中值濾波器、最鄰近均值濾波器、最鄰近中值濾波器、調適濾波器、高斯濾波器、桑原(Kuwahara)濾波器以及鈍化遮罩)處理。 實驗一以相似度分析進行評估，目的在於利用XS3影像與PAN影像之差異以驗證主軸轉換(eigenspace projection)為可用之評估方式之一。實驗二則採取四種評估機制:傳統數值指標(traditional numerical indices)、客觀評估(subjective criteria)、銳化/平滑化指標(sharpening/smoothing measure)以及相似度分析(similarity/distance measure)進行評估，目的在於比較四種方式之優劣。

For mapping or object identification, the edges possess important information. It would be desirable to preserve the edges in the original image, while applying smoothing filter to reduce the influence of noise. A number of filters are available for this purpose, including the Minimum Coefficient of Variation (MCV) Filter, Kuwahara Filter, Nagao Filter, SNNF (Symmetric Nearest Neighbor Filter). This study investigates the evaluation scheme for the performance of these filters. Traditional numerical indices, similarity/distance measures, subjective criteria method and the recently developed Smoothing/Sharpening measures are applied for comparison. From the experiments with simulated noises, it is found that the Smoothing/Sharpening measures do provide intrinsic information. However, the meaning of the quality still remains for discussion.