A way to convert YUV (NV12) files to RGB (BMP). Requires PIL (Python Image Library)

yuv2rgb.py

__author__ = "Farid Zakaria"
__copyright__ = "Copyright 2011, AMD"
__credits__ = ["Sherin Sasidhan ([email protected])"]
__license__ = "GPL"
__version__ = "1.0."
__maintainer__ = "Farid Zakaria"
__email__ = "[email protected]"
__status__ = "Prototype"

import argparse
from PIL import Image
import sys
from struct import *
import os
import math

DEFAULT_STRIDE = -1
DEFAULT_FRAME = 0

parser = argparse.ArgumentParser(prog='YUV2RGB Converter', description='Convert YUV Raw images into RGB format')
parser.add_argument('filename', type=str)
parser.add_argument('width', type=int)
parser.add_argument('height', type=int)
parser.add_argument('--stride', default=DEFAULT_STRIDE, type=int, help='stride of the image (default: width of the image')
parser.add_argument('--frame', default=DEFAULT_FRAME, type=int, help='frame number to grab (default: index 0)')
parser.add_argument('--version',action='version', version='%(prog)s 1.0')

class Converter(object):
	"""Base class that should define interface for all conversions"""
	def __init__(self, filename, width, height, stride, frame):
		self.filename = filename
		self.width = width
		self.height = height
		self.stride = stride
		self.frame = frame
	#constructor#

	def Convert():
		raise NotImplementedError( "Should have implemented this!" )
	#convert#
#Converter#



class NV12Converter(Converter):
	"""This class converts NV12 files into RGB"""
	def __init__(self, filename, width, height, stride, frame):
		super(NV12Converter, self).__init__(filename, width, height, stride, frame)
	#constructor#

	def Convert(self):
		f_y = open(self.filename, "rb")
		f_uv= open(self.filename, "rb")

		converted_image_filename = self.filename.split('.')[0] + ".bmp"
		converted_image = Image.new("RGB", (self.width, self.height) )
		pixels = converted_image.load()

		size_of_file = os.path.getsize(self.filename)
		size_of_frame = ((3.0/2.0)*self.height*self.width)
		number_of_frames = size_of_file / size_of_frame
		frame_start = size_of_frame * self.frame
		uv_start = frame_start + (self.width*self.height)
        
 		#lets get our y cursor ready
		f_y.seek(frame_start);        
		for j in range(0, self.height):
			for i in range(0, self.width):
				#uv_index starts at the end of the yframe.  The UV is 1/2 height so we multiply it by j/2
				#We need to floor i/2 to get the start of the UV byte
				uv_index = uv_start + (self.width * math.floor(j/2)) + (math.floor(i/2))*2
				f_uv.seek(uv_index)
				
				y = ord(f_y.read(1))
				u = ord(f_uv.read(1))
				v = ord(f_uv.read(1))
				
				b = 1.164 * (y-16) + 2.018 * (u - 128)
				g = 1.164 * (y-16) - 0.813 * (v - 128) - 0.391 * (u - 128)
				r = 1.164 * (y-16) + 1.596*(v - 128)
	
				pixels[i,j] = int(r), int(g), int(b)


		converted_image.save(converted_image_filename)




def main():
	args = parser.parse_args()

	if args.stride == DEFAULT_STRIDE:
		args.stride = args.width

	if args.frame < 0:
		args.frame = 0

	converter = NV12Converter(args.filename, args.width, args.height, args.stride, args.frame)
	converter.Convert()

#main function#

if __name__ == "__main__":
	main()

Thank you, it works. But why is it not natural as the original?

not sure the question

Sorry, because of different screens the intensivity of the photo was different.