0x00. 使用 Canny 算法边缘识别
Canny 算法是一种多级边缘识别算法。
Canny边缘识别算法可以分为以下5个步骤:
-
应用高斯滤波来平滑图像,目的是去除噪声。
-
找寻图像的强度梯度(intensity gradients)。
-
应用非最大抑制(non-maximum suppression)技术来消除边误检(本来不是但检测出来是)。
-
应用双阈值的方法来决定可能的(潜在的)边界。
-
利用滞后技术来跟踪边界。
具体原理性质的东西可以参考这里
读取本地视频处理代码示例:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 |
import cv2.cv as cv capture = cv.CaptureFromFile('img/myvideo.avi') nbFrames = int(cv.GetCaptureProperty(capture, cv.CV_CAP_PROP_FRAME_COUNT)) fps = cv.GetCaptureProperty(capture, cv.CV_CAP_PROP_FPS) wait = int(1/fps * 1000/1) dst = cv.CreateImage((int(cv.GetCaptureProperty(capture, cv.CV_CAP_PROP_FRAME_WIDTH)), int(cv.GetCaptureProperty(capture, cv.CV_CAP_PROP_FRAME_HEIGHT))), 8, 1) for f in xrange( nbFrames ): frame = cv.QueryFrame(capture) cv.CvtColor(frame, dst, cv.CV_BGR2GRAY) cv.Canny(dst, dst, 125, 350) cv.Threshold(dst, dst, 128, 255, cv.CV_THRESH_BINARY_INV) cv.ShowImage("The Video", frame) cv.ShowImage("The Dst", dst) cv.WaitKey(wait) |
直接处理摄像头视频:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 |
import cv2.cv as cv capture = cv.CaptureFromCAM(0) dst = cv.CreateImage((int(cv.GetCaptureProperty(capture, cv.CV_CAP_PROP_FRAME_WIDTH)), int(cv.GetCaptureProperty(capture, cv.CV_CAP_PROP_FRAME_HEIGHT))), 8, 1) while True: frame = cv.QueryFrame(capture) cv.CvtColor(frame, dst, cv.CV_BGR2GRAY) cv.Canny(dst, dst, 125, 350) cv.Threshold(dst, dst, 128, 255, cv.CV_THRESH_BINARY_INV) cv.ShowImage("The Video", frame) cv.ShowImage("The Dst", dst) c = cv.WaitKey(1) if c == 27: #Esc on Windows break |
0x01. 人脸识别
使用OpenCV可以很简单的检测出视频中的人脸等:
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 |
import cv2.cv as cv capture=cv.CaptureFromCAM(0) hc = cv.Load("haarcascades/haarcascade_frontalface_alt.xml") while True: frame=cv.QueryFrame(capture) faces = cv.HaarDetectObjects(frame, hc, cv.CreateMemStorage(), 1.2,2, cv.CV_HAAR_DO_CANNY_PRUNING, (0,0) ) for ((x,y,w,h),stub) in faces: cv.Rectangle(frame,(int(x),int(y)),(int(x)+w,int(y)+h),(0,255,0),2,0) cv.ShowImage("Window",frame) c=cv.WaitKey(1) if c==27 or c == 1048603: #If Esc entered break |