使用Python+PIL查看两张相似图形的差异

PIL(http://www.pythonware.com/products/pil/)中包含很多图形处理库，主要是Image，Image库中包含三个类（Image、ImagePointHandler、ImageTransformHandler）及若干方法。

今天实验了一把，打开两个差异很小的图片，将其中一个反色处理，然后在将两个图像混合，就能比较清晰的看到差异的地方。

唔~~~~，这要是扩展一下，增加截图功能，可以做成“大家来找茬”之类游戏的作弊器了！！阿门~~~

代码如下：

import Image, ImageChops
im1 = Image.open('c:/1.bmp')
im2 = Image.open('c:/2.bmp')
im3 = ImageChops.invert(im2)
Image.blend(im1,im3,0.5).show()

原始图片如下：

1.bmp



2.bmp



混合后结果：



附：PIL.Image库说明文档

PIL.Image

# 2009-11-15 fl PIL release 1.1.7

Classes
		ImageImagePointHandlerImageTransformHandler class Image Methods defined here: __getattr__(self, name)__init__(self)__repr__(self)convert(self, mode=None, data=None, dither=None, palette=0, colors=256)Convert to other pixel formatcopy(self) Copy raster data复制一个新的对象crop(self, box=None)Crop region from imagedraft(self, mode, size) Configure image decoderfilter(self, filter) Apply environment filter to imagefromstring(self, data, decoder_name='raw', *args)Load data to image from binary string从原始图像数据中加载图像，一般用来从剪切板中获取图像getbands(self) Get band namesgetbbox(self) Get bounding box of actual data (non-zero pixels) in image获取当前图像在实际图片中的范围getcolors(self, maxcolors=256)Get colors from image, up to given limitgetdata(self, band=None)Get image data as sequence object.获取图像数据列表。如果指定band，就获取这个颜色通道的值得列表 print im.getdata()[0] (253,23,34) getextrema(self) Get min/max valuegetim(self) Get PyCObject pointer to internal image memorygetpalette(self) Get palette contents.getpixel(self, xy) Get pixel value获取每个坐标像素点的RGB值。参数xy格式为(x,y)getprojection(self) Get projection to x and y axeshistogram(self, mask=None, extrema=None)Take histogram of imageload(self) Explicitly load pixel data.offset(self, xoffset, yoffset=None)(deprecated) Offset image in horizontal and/or vertical directionpaste(self, im, box=None, mask=None)Paste other image into regionpoint(self, lut, mode=None)Map image through lookup tableputalpha(self, alpha) Set alpha layerputdata(self, data, scale=1.0, offset=0.0)Put data from a sequence object into an image.putpalette(self, data, rawmode='RGB')Put palette data into an image.putpixel(self, xy, value) Set pixel value设置某坐标像素点值quantize(self, colors=256, method=0, kmeans=0, palette=None)resize(self, size, resample=0)Resize image重新设置图像大小rotate(self, angle, resample=0, expand=0)Rotate image. Angle given as degrees counter-clockwise.旋转图像save(self, fp, format=None, **params)Save image to file or stream保存图像seek(self, frame) Seek to given frame in sequence file定位到动画某一帧，一般用于gif格式的文件图片show(self, title=None, command=None)Display image (for debug purposes only)现实图像，调用默认的图像浏览器split(self) Split image into bands将图像按照RGB分别解析，返回三个Image对象tell(self) Return current frame number返回当前帧号thumbnail(self, size, resample=0)Create thumbnail representation (modifies image in place)创建缩略图，主要用来修改图像大小和清晰度，以便生产较小的文件tobitmap(self, name='image')Return image as an XBM bitmaptostring(self, encoder_name='raw', *args)Return image as a binary stringtransform(self, size, method, data=None, resample=0, fill=1)Transform imagetranspose(self, method) Transpose image (flip or rotate in 90 degree steps)verify(self) Verify file contents. Data and other attributes defined here: format = None mode format_description = None classImagePointHandler classImageTransformHandler

Functions
		blend(im1, im2, alpha) Interpolate between images.将两张图片混合到一起。参数alpha是im1与im2的混合权重，范围为[0,1] composite(image1, image2, mask)Create composite image by blending images using a transparency mask eval(image, *args) Evaluate image expression fromarray(obj, mode=None) frombuffer(mode, size, data, decoder_name='raw', *args)Load image from string or buffer fromstring(mode, size, data, decoder_name='raw', *args)Load image from string getmodebandnames(mode) getmodebands(mode) getmodebase(mode) getmodetype(mode) init() Load all file format drivers. isDirectory(f) isImageType(t) isNumberType(...) isNumberType(a) -- Return True if a has a numeric type, False otherwise. isSequenceType(...) isSequenceType(a) -- Return True if a has a sequence type, False otherwise. isStringType(t) ## # (Internal) Checks if an object is a string. If the current # Python version supports Unicode, this checks for both 8-bit # and Unicode strings. isTupleType(t) merge(mode, bands) Merge a set of single band images into a new multiband image. new(mode, size, color=0)Create a new image创建一个Image对象 open(fp, mode='r')Open an image file, without loading the raster data打开图片文件 preinit() Load standard file format drivers. register_extension(id, extension) register_mime(id, mimetype) register_open(id, factory, accept=None) register_save(id, driver)