Computer Vision Projects

The followings are project set from Advance Computer Vision class, all projects use OpenCV Image I/O API only. The rest of algorithms are implemented by me.

以下為高等電腦圖學有趣的作品集,全部的內容僅使用 OpenCV 讀寫圖檔外,其餘的演算法皆自行實作。


Origin Pic Lena/ Lena 原圖



Gaussian Blur/高斯模糊




Corner Detection/角點偵測





Edge Detection/邊緣偵測




Train Identification/火車辨識

In this project, I use OpenCV to split a image into 3x3 sub images, and calculate the central moments of each sub image. Some images are flipped, noised, or edge strengthen due to the lack of image to create a good model. These results are sent to SVM ( Support Vector Machine ) to build the model. After that, we can take any image as input to see which train it is. The shape and texture detections should be future enhancement to increase the accuracy.

此專案使用 OpenCV 將一張照片切割成 3x3 的子照片,再對每一個子照片計算其 Central Moments 。來源照片數量有限所以經過翻轉、雜訊、邊緣強化等增加母體照片數。這些計算結果送入 SVM 建立模型,之後便可將測試照片輸入此模型,由系統判斷為何種車種。未來應考慮形狀和紋理辨識增加準確度。