OpenCV Computer Vision with Python

这是一篇读书笔记。

Setting up OpenCV

Related Libs

NumPy, SciPy, OpenNI (optional), SenorKinect (optional)

Different OS

Sample codes

Handing Files, Cameras, and GUIs

Basic I/O Scripts

Reading/Writing an Image file

import cv2
image = cv2.imread('/data/pic/lena.jpg')
image_gray = cv2.imread('/data/pic/lena.jpg', cv2.CV_LOAD_IMAGE_GRAYSCALE)
cv2.imwrite('Lena.png', image)
cv2.imwrite('Lena_gray.png', image_gray)

Converting between an image and raw bytes

import cv2, numpy, os
randomByteArray = bytearray(os.urandom(120000))
flatNumpyArray = numpy.array(randomByteArray)
grayImage = flatNumpyArray.reshape(300, 400)
cv2.imwrite('RandomGray.png', grayImage)
bgrImage = flatNumpyArray.reshape(100, 400, 3)
cv2.imwrite('RandomBGR.png', bgrImage)

Reading/Writing a video file

import cv2
videoCapture = cv2.VideoCapture('/data/ac-demo.avi')
fps       = videoCapture.get(cv2.cv.CV_CAP_PROP_FPS)
size = (int(videoCapture.get(cv2.cv.CV_CAP_PROP_FRAME_WIDTH)),
        int(videoCapture.get(cv2.cv.CV_CAP_PROP_FRAME_HEIGHT)))
videoWriter = cv2.VideoWriter('MyOutputVideo.avi',
                              cv2.cv.CV_FOURCC('I', '4', '2', '0'),
                              fps,
                              size)
success, frame = videoCapture.read()
while success: # Loop until there are no more frames
    videoWriter.write(frame)
    success, frame = videoCapture.read()

# 这段代码我失败了……
cv2.cv.CV_FOURCC('T','H','E','O'): Ogg-Vorbis, *.ogv
cv2.cv.CV_FOURCC('F','L','V','1'): Flash video, *.flv
cv2.cv.CV_FOURCC('M','J','P','G'): motion-JPEG, *.avi
cv2.cv.CV_FOURCC('P','I','M','1'): MPEG-1, *.avi
cv2.cv.CV_FOURCC('I','4','2','0'): uncompressed YUV, *.avi

Capturing camera frames

import cv2
cameraCapture = cv2.VideoCapture(0)
fps = 30 # an assuption
size = (int(cameraCapture.get(cv2.cv.CV_CAP_PROP_FRAME_WIDTH)),
        int(cameraCapture.get(cv2.cv.CV_CAP_PROP_FRAME_HEIGHT)))
print "size: ", size
videoWriter = cv2.VideoWriter('MyOutputVid.avi',
                               cv2.cv.CV_FOURCC('M','J','P','G'),fps, size)
success, frame = cameraCapture.read()
numFramesRemaining = 3 * fps - 1
while success and numFramesRemaining > 0:
    videoWriter.write(frame)
    success, frame = cameraCapture.read()
    numFramesRemaining -= 1
    print str(numFramesRemaining)
# 再次失败

Displaying camera frames in a window

import cv2
clicked = False
def onMouse(event, x, y, flags, param):
    global clicked
    if event == cv2.cv.CV_EVENT_LBUTTONUP:
        clicked = True
cameraCapture = cv2.VideoCapture(0)
cv2.namedWindow('MyWindow')
cv2.setMouseCallback('MyWindow', onMouse)
print 'Showing camera feed. Click window or press any key to stop.'
success, frame = cameraCapture.read()
while success and cv2.waitKey(1) == -1 and not clicked:
    cv2.imshow('MyWindow', frame)
    success, frame = cameraCapture.read()
cv2.destroyWindow('MyWindow')
# 终于出来一次……

cv2.cv.CV_EVENT_

              MOUSEMOVE
LBUTTONUP     MBUTTONUP     RBUTTONUP
LBUTTONDOWN   MBUTTONDOWN   RBUTTONDOWN
LBUTTONDBCLK  MBUTTONDBCLK  RBUTTONDBCLK

Project concept

We will develop an interactive application that performs face tracking and image, and manipulations on camera input in real time.

extensible and reusable

An object-oriented design

Python applications can be written in a purely procedural style. This is often done with small applications like our basic I/O scripts, discussed previously. However, from now on, we will use an object-oriented style because it promotes modularity and extensibility.

Python Knowledge

Filtering Images

  1. Creating modules

reuseable modules: filters.py, utils.py

Channel mixing – seeing in Technicolor

channel mixing

remapping colors to simulate some other, smaller color space inside RGB or BGR

Color Spaces

Simulating RC color space

# Pseudocode:
#     dst.b = dst.g = 0.5 * (src.b + src.g)
#     dst.r = src.r
# Python OpenCV Code:
b, g, r = cv2.split(src)
cv2.addWeighted(b, 0.5, g, 0.5, 0, b)
cv2.merge((b, b, r), dst)

Simulating RGV color space

# Pseudocode:
#     dst.b = min(src.b, src.g, src.r),
#     dst.g = src.g,
#     dst.r = src.r
# Python code:
b, g, r = cv2.split(src)
cv2.min(b, g, b)
cv2.min(b, r, b)
cv2.merge((b, g, r), dst)

Simulating CMV color space

# Pseudocode:
#     dst.b = max(src.b, src.g, src.r),
#     dst.g = src.g,
#     dst.r = src.r
Curves – bending color space

Curves are another technique for remapping colors.

 dst.r = funcR(src.r)
 dst.g = funcG(src.g)
 dst.b = funcB(src.b)

cubic spline interpolation, or

Formulating a curve

def createCurveFunc(points):
   """Return a function derived from control points."""
   if points is None:
       return None
   numPoints = len(points)
   if numPoints < 2:
       return None
   xs, ys = zip(*points)
   if numPoints < 4:
       kind = 'linear'
       # 'quadratic' is not implemented.
   else:
       kind = 'cubic'
   return scipy.interpolate.interp1d(xs, ys, kind, bounds_error = False)

Caching and applying a curve

in utils.py

# def createLookupArray(func, length = 256)
#     min(max(0, func_i), length - 1) # ensure fun_i » [0, length - 1]
# def applyLookupArray(lookupArray, src, dst)
#     dst[:] = lookupArray[src]
def createCompositeFunc(func0, func1):
        """Return a composite of two functions."""
        if func0 is None:
            return func1
        if func1 is None:
            return func0
        return lambda x: func0(func1(x))
# Misc:
flatView = array.view()
flatView.shape = array.size
# numpy.view only owns a reference to the data, not a copy

Designing object-oriented curve filters

Emulating photo films

Highlighting edges

cv2.medianBlur(src, blurKsize)
cv2.Laplacian(graySrc, cv2.cv.CV_8U, graySrc, ksize = edgeKsize)
normalizedInverseAlpha = (1.0 / 255) * (255 - graySrc)
channels = cv2.split(src)
for channel in channels:
    channel[:] = channel * normalizedInverseAlpha
cv2.merge(channels, dst)

Custom kernels – getting convoluted

kernel

convolution matrix, it mixes up or convolutes the pixels in a region.

  kernel = numpy.array([[-1, -1, -1],
                        [-1,  9, -1],
                        [-1, -1, -1]])

Tracking Faces with Haar Cascades

Conceptualizing Haar cascades

image data \(\to\) features (e.g. Haar-like features) \(\to\) cascade (collection of features)

cascade

Two images that differ only in scale should be capable of yielding similar features, albeit for different window sizes. Thus, it is useful to generate features for multiple window sizes. Such a collection of features is called a cascade.

a Haar cascade is scale-invariant or, in other words, robust to changes in scale

Getting Haar cascade data

Defining a face as a hierarchy of rectangles

class Face(object):
    """Data on facial features: face, eyes, nose, mouth."""
    def __init__(self):
        self.faceRect = None
        self.leftEyeRect = None
        self.rightEyeRect = None
        self.noseRect = None
        self.mouthRect = None

Tracing, cutting, and pasting rectangles

def outlineRect(image, rect, color):
    if rect is None:
        return
    x, y, w, h = rect
    cv2.rectangle(image, (x, y), (x+w, y+h), color)
cv2.INTER_NEAREST
cv2.INTER_LINEAR # bilinear interpolation (the default)
cv2.INTER_AREA
cv2.INTER_CUBIC # bicubic interpolation over a 4 x 4 pixel neighborhood
cv2.INTER_LANCZOS4 # This is Lanczos interpolation over an 8 x 8 pixel neighborhood, the highest-cost, highest-quality approach

\[L(x) = \begin{cases} \operatorname{sinc}(x)\, \operatorname{sinc}(x/a) & \text{if}\;\; -a < x < a\\ 0 & \text{otherwise} \end{cases}\]

Lanczos windows for a = 1, 2, 3.

Lanczos windows for a = 1, 2, 3.

Equivalently,

\[L(x) = \begin{cases} 1 & \text{if}\;\; x = 0\\ \quad & \\ \displaystyle \frac{a \sin(\pi x) \sin(\pi x / a)}{\pi^2 x^2} & \text{if}\;\; 0 < |x| < a \\ \quad & \\ 0 & \text{otherwise} \end{cases}\]

Adding more utility functions

def isGray(image):
    """Return True if the image has one channel per pixel."""
    return image.ndim < 3
def widthHeightDividedBy(image, divisor):
    """Return an image's dimensions, divided by a value."""
    h, w = image.shape[:2]
    return (w/divisor, h/divisor)

Detecting Foreground/Background Regions and Depth

dvorak4tzx:

I dont have Kinect camera, so pass this chapter, and that’s all the book covers

Appendix A: Integrating with Pygame

Appendix B: Generating Haar Cascades for Custom Targets

Refs

  1. OpenCV Computer Vision with Python
  2. Joseph Howse(the author)’s webpage
  3. Lanczos resampling - Wikipedia, the free encyclopedia