摘要
中文摘要 要構成一張彩色影像需要紅、藍、綠三種顏色，所以為了感測三種顏色的資訊，必須使用三片感光元件來感測，但是卻會使的相機的成本提高，因此現在許多消費型相機為了減少成本而採用單一感光元件來感測，所以每個像素只能儲存一個顏色資訊，因此代表著每個像素都遺失掉兩個顏色值，所以我們必須藉由鄰近的已知資訊來估測遺失的顏色值，這估計的方法我們稱為顏色補插演算法，在第二章我們將介紹六種補插的演算法並分析之間的優缺點，讓大家對補插演算法有所了解。 目前有許多顏色補插的方式，但最終的目的都是要讓補插後的影像能夠達到最好的品質與PSNR，但是因為影像本身的色彩分佈與特性都不一樣，變化很多，所以如果能夠找到一個針對大部分影像都能達到最好的補插結果，那是我們的目標之一，另一個重點就是補插時的運算複雜度，目前要達到比較好的影像品質，通常都是需要經過比較複雜的運算，在硬體設計上也會比較煩瑣，如果在補插後的影像能達到一樣的品質，但是在運算量上卻能夠簡化，在補插演算法上也是一大改進。在第三章除了介紹我們的補插方法並與第二章所介紹的演算法一起比較，並針對運算複雜度和補插後的影像品質去分析比較，讓大家更能夠了解我們的演算法與其他方法之間的優、缺點。
關鍵字
補差並列摘要
Abstract To compose a colorful image needs three colors: red, blue and green. In order to sense the information of the three colors, we have to use 3-CCD to sense it, but that will increase the cost of cameras. Therefore, there are many consuming cameras using single CCD to sense to decrease cost, so each pixel can only save the information of one color. As a result, it represents that each pixel loses two color values at the same time. That’s why we need to estimate the missing color values by close and known information. The estimative method is called Interpolation Algorithm. There are a lot of interpolation methods, and all the destination is to make interpolated image accomplishes best quality and PSNR. However, color spreading and characters of image itself are different and various. One of our goals is to find the best interpolative method to perfect most of the images. And another point is the operation complexity when we interpolate. At present, we can say that the better image quality you need, the more complicated operation you have. It is also more trifling on hardware design as well. If we can reach the same quality after interpolating but with easier operation, it is also a great improvement on Interpolation Algorithm.