Open CV Structure CV MLL High GUI Image

Open. CV Structure CV MLL High. GUI Image Processing and Vision Algorithms Statistical Classifiers and Clustering Tools GUI, Image and Video I/O CXCORE basic structures and algorithms, XML support, drawing functions IPP Fast architecture-specific low-level functions Aux Gary or Bradski, Experimental less 2009 used routines 10 10

存取像素 http: //www. cs. iit. edu/~agam/cs 512/lect-notes/opencv-intro. html#SECTION 0005300000000

Frame Difference 程式碼

混合高斯背景模型 // initialize Mo. G parameters int total. Gaussian = 5; 想模擬的分布數量 int max. Gaussian = 5; 找尋前景時，只找前max. Gaussian個依weight大小排序後的分布 float std. Deviation. Init = 6; 預設的標準差 float std. Deviation. Threshold = 3. 0; 判斷前景的標準(多少倍標準差之外) float weight. Threshold = 0. 25; 找尋前景時，只找weight高於此標準的分布 float alpha = 0. 01; 背景學習率，愈大背景改變愈快，愈小愈有容忍度。

混合高斯背景模型 Please try our solution – Gaussian. Background. Model 目前total. Gaussian, max. Gaussian一定要 一樣 – 嘗試實作將分布排序，然後可以設定只考慮依 weight排序後的前max. Gaussian個分布

動態歷史圖像(Motion History Image, MHI)

範例 – 請參照原始碼 – motempl. c – Or our solution in Motion. History. Image – 一些變數 – MHI_DURATION – 想要追蹤的時間長度(secs) – N – 想要計算motion detection的Frame buffer長度(frames) Challenge -- 嘗試將gradient圖標記拿掉

How to use Try our package – Face. Detection – Face. Detect. cmd Usage: facedetect [-cascade="<cascade_path>"] [--nestedcascade[="nested_cascade_path"]] [--scale[=<image scale> [filename|camera_index] – Face. Detect. exe 執行辨識的程式

前景臉部偵測 Facedetect --cascade=". /haarcascades/haarcascade_frontalface_alt. xml” --nested-cascade=". /haarcascades/haarcascade_eye. xml” --scale=1. 3 先找到在前面的臉 接者尋找眼睛 縮放標記

試試其他模組 haarcascade_eye. xml haarcascade_eye_tree_eyeglasses. xml haarcascade_frontalface_alt 2. xml haarcascade_frontalface_alt_tree. xml haarcascade_frontalface_default. xml haarcascade_fullbody. xml haarcascade_lefteye_2 splits. xml haarcascade_lowerbody. xml haarcascade_mcs_eyepair_big. xml haarcascade_mcs_eyepair_small. xml haarcascade_mcs_lefteye. xml haarcascade_mcs_mouth. xml haarcascade_mcs_nose. xml haarcascade_mcs_righteye. xml haarcascade_mcs_upperbody. xml haarcascade_profileface. xml haarcascade_righteye_2 splits. xml haarcascade_upperbody. xml

如何製作自己的Xml 你必須收集一些樣品去檢測… 接著校導… 產生 xml 檔 接著測試… Please refer to following link: – http: //note. sonots. com/Sci. Software/haartra ining. html

推薦的參考網站 Open Computer Vision Library (Sourceforge) – http: //sourceforge. net/projects/opencvlibrary/ Open. CV Official Forum – http: //tech. groups. yahoo. com/group/Open. CV/ Open. CV Wiki – http: //opencv. willowgarage. com/wiki/ (Willowgarage) – http: //en. wikipedia. org/wiki/Open. CV (Wikipedia) Open. CV中文網站 – http: //www. opencv. org. cn/index. php/ 優質Open. CV教學網 – http: //yester-place. blogspot. com/ Learning Open. CV: Computer Vision with the Open. CV Library (Paperback) – http: //www. amazon. com/Learning-Open. CV-Computer-Vision. Library/dp/0596516134

參考 http: //vbie. eic. nctu. edu. tw/vol_2/skill_7. htm

參考 http: //vbie. eic. nctu. edu. tw/vol_2/skill_7. htm http: //vbie. eic. nctu. edu. tw/vol_13/tech 1. htm

參考 http: //www. cse. ohiostate. edu/~jwdavis/CVL/Research/MHI/ mhi. html

參考 特徵檢測專題 – http: //www. opencv. org. cn/index. php/%E 7 %89%B 9%E 5%BE%81%E 6%A 3%80%E 6%B 5 %8 B%E 4%B 8%93%E 9%A 2%98