python PIL 图像处理模块以及array的互相转换,基本图像处理收集

http://pillow.readthedocs.org/en/3.1.x/reference/Image.html

http://blog.csdn.net/xueweuchen/article/details/38756075

http://python-reference.readthedocs.org/en/latest/docs/functions/bytearray.html

https://github.com/yipinp/ADAS_Lab

PIL 库中比较好用的就是fromarray等。

PIL.Image. fromarray ( obj,  mode=None )

 

Creates an image memory from an object exporting the array interface (using the buffer protocol).

If obj is not contiguous, then the tobytes method is called and frombuffer() is used.

Parameters:	obj – Object with array interface mode – Mode to use (will be determined from type if None) See: Modes.
Returns:	An image object.

PIL.Image. frombytes ( mode,  size,  data,  decoder_name='raw',  *args )

Creates a copy of an image memory from pixel data in a buffer.

In its simplest form, this function takes three arguments (mode, size, and unpacked pixel data).

You can also use any pixel decoder supported by PIL. For more information on available decoders, see the section Writing Your Own File Decoder.

Note that this function decodes pixel data only, not entire images. If you have an entire image in a string, wrap it in a BytesIO object, and use open() to load it.

Parameters:	mode – The image mode. See: Modes. size – The image size. data – A byte buffer containing raw data for the given mode. decoder_name – What decoder to use. args – Additional parameters for the given decoder.
Returns:	An Image object.

在python中本身还有bytearray

bytearray([source [, encoding [, errors]]])

中文说明：

bytearray([source [, encoding [, errors]]])返回一个byte数组。Bytearray类型是一个可变的序列，并且序列中的元素的取值范围为 [0 ,255]。

参数source:

如果source为整数，则返回一个长度为source的初始化数组；

如果source为字符串，则按照指定的encoding将字符串转换为字节序列；

如果source为可迭代类型，则元素必须为[0 ,255]中的整数；

如果source为与buffer接口一致的对象，则此对象也可以被用于初始化bytearray.。

版本：在python2.6后新引入，在python3中同样可以使用！

Conversions between Various Formats
	Convert between Python tuple and list a = (1, 2)      # a is a tuple  b = list(a)     # b is a list  c = tuple(b)  # c is a tuple  Convert between Python tuple, list and NumPy 1D array  a = (1, 2)          # a is a tuple  b = np.array(a)  # b is an numpy array  c = tuple(b)       # c is a tuple  a = [1, 2]          # a is a python array  b = np.array(a)  # b is a numpy array  c = list(b)          # c is a python list  Convert between NumPy 2D array and NumPy matrix a = numpy.ndarray([2,3])   # create 2x3 array m1 = numpy.asmatrix(a)   # does not create new matrix, m1 refers to the same memory as a  m2 = numpy.matrix(a)      # creates new matrix and copies content  b1 = numpy.asarray(m2)   # does not create array, b1 refers to the same memory as m2 b2 = numpy.array(m2)      # creates new array and copies content
	Read/Write Images with OpenCV image = cv.LoadImage(“ponzo.jpg”) cv.SaveImage(“out.png”, image) Read/Write Images with PIL image = Image.open(“ponzo.jpg”) image.save(“out.jpg”) Read/Write Images with PyOpenCV mat = pyopencv.imread(“ponzo.jpg”) pyopencv.imwrite(“out.png”, mat)
	Convert between OpenCV image and PIL image # color image  cimg = cv.LoadImage("ponzo.jpg", cv.CV_LOAD_IMAGE_COLOR)     # cimg is a OpenCV image pimg = Image.fromstring("RGB", cv.GetSize(cimg), cimg.tostring())    # pimg is a PIL image  cimg2 = cv.CreateImageHeader(pimg.size, cv.IPL_DEPTH_8U, 3)      # cimg2 is a OpenCV image  cv.SetData(cimg2, pimg.tostring()) Note: OpenCV stores color image in BGR format. So, the converted PIL image is also in BGR-format. The standard PIL image is stored in RGB format.  # gray image  cimg = cv.LoadImage("ponzo.jpg", cv.CV_LOAD_IMAGE_GRAYSCALE)   # cimg is a OpenCV image  pimg = Image.fromstring("L", cv.GetSize(cimg), cimg.tostring())                 # pimg is a PIL image  cimg2 = cv.CreateImageHeader(pimg.size, cv.IPL_DEPTH_8U, 1)              # cimg2 is a OpenCV image cv.SetData(cimg2, pimg.tostring())
	Convert between PIL image and NumPy ndarray image = Image.open(“ponzo.jpg”)   # image is a PIL image  array = numpy.array(image)          # array is a numpy array  image2 = Image.fromarray(array)   # image2 is a PIL image  Convert between PIL image and PyOpenCV matrix image = Image.open(“ponzo.jpg”)                  # image is a PIL image mat = pyopencv.Mat.from_pil_image(image)  # mat is a PyOpenCV matrix  image2 = mat.to_pil_image()                        # image2 is a PIL image
	Convert between OpenCV image and NumPy ndarray cimg = cv.LoadImage("ponzo.jpg", cv.CV_LOAD_IMAGE_COLOR)   # cimg is a OpenCV image  pimg = Image.fromstring("RGB", cv.GetSize(cimg), cimg.tostring())  # pimg is a PIL image  array = numpy.array(pimg)     # array is a numpy array  pimg2 = cv.fromarray(array)    # pimg2 is a OpenCV image

-------------------------------------------------------------------------------------------------------------

1. opencv   videoWriter写视频文件

# Create video writer object vidwrite = cv2.VideoWriter(['testvideo',cv2.cv.CV_FOURCC('M','J','P','G'), 25, (640,480),True])

注意到如果要写成功有几个因素需要注意：

  （1） 必须安装了ffmpeg或者Gstreamer

  （2） 摄像头的配置参数必须和写文件的参数一致,注意的是width，height，是width在前

      cap = cv2.VideoCapture(0)

      cap.set(1,20.0)

      cap.set(3,640)

      cap.set(4,480)

      fourcc = cv2.cv.FOURCC('M','J','P','G')

      videoW = cv2.VideoWriter('d.avi',fourcc,20.0,(640,480),True)

      while(True):

            ret,frame = cap.read()

            cv2.imshow('frame',frame)

            videoW.write(frame)

            if cv2.waitKey(1) & 0xff == ord('q'):

                  break

      videoW.release() 

      cap.release()

      cv2.destroyAllWindows()

也可以使用：

http://editcode.info/questions/26452909/opencv-write-frame-to-file-python

# The following might fail if the device doesn't support these values
cap.set(1, 20.0) #Match fps
cap.set(3,640)   #Match width
cap.set(4,480)   #Match height

# So it's always safer to retrieve it afterwards
fps = cap.get(CV_CAP_PROP_FPS)
w = cap.get(CV_CAP_PROP_FRAME_WIDTH);
h = cap.get(CV_CAP_PROP_FRAME_HEIGHT);

2. opencv画直线等

注意到这些函数都是不支持分数像素精度的，只是支持整数point的输入。