用頻率分析降低影像雜訊

Image Noise Reduction with Frequency Analysis

鍾招佑(Chou-Yuh Chong)

指導教授 ： 傅楸善

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2008年)

https://doi.org/10.6342/NTU.2008.01126

摘要

數位相機拍照的影像，因為光學上的原因和儀器本身的限制，所形成的影像會有一定的雜訊。為了達到可以讓消費者接受的影像品質，數位相機的數位訊號處理器必須包含降低雜訊的功能。
如果使用相機裡的數位訊號處理器降低雜訊，會比後處理的降低雜訊的方式還要多一點限制。比如計算效能的限制，還有記憶體限制，甚至在時效上也需要一定程度的限制。本方法即是考慮了這些限制下，盡量簡化演算法，避免疊代和大量記憶體空間的方式，盡量分析區域內不同頻率的資訊，達到降低雜訊的效果。

關鍵字

影像 ； 雜訊 ； 頻率分析

並列摘要

Images captured from digital still cameras are often with a certain amount of noise which is subjected to optical sources and physical limitation of instruments. To capture image with quality acceptable by customers, image signal processor in digital still camera should include noise reduction.

If we use image signal processor in digital still camera, rather than post-processing software on personal computer, to reduce noise, we must consider some limitations, such as computation power, memory limit, or even some time constraint. Our method does consider these limitations, making our algorithm as simple as possible, preventing any iterative repetition, and avoiding huge memory usage. We also analyze as much information as possible locally in different frequencies to reduce noise.

並列關鍵字

Image ； Noise Reduction ； Frequency Analysis

參考文獻

[1] A. Petrosyan and A. Ghazaryan, “Method and System for Digital Image Enhancement,” US Patent# 11/116408, 2006.

[3] Digital Photography Tutorial, “Digital Photography Tutorial," http://www.cambridgeincolour.com/tutorials.htm, 2008.

[5] PH. Mei, “An Introduction to Digital Camera Signal Processor”, 2006

[2] Curvelet, “Curvelet.Org,” http://www.wavelet.org/, 2008.

[4] Nikon, “Nikon D70”, http://www.nikondigitutor.com/eng/d70/index.php, 2008