matplotlib.image#

The image module supports basic image loading, rescaling and displayoperations.

classmatplotlib.image.AxesImage(ax,*,cmap=None,norm=None,colorizer=None,interpolation=None,origin=None,extent=None,filternorm=True,filterrad=4.0,resample=False,interpolation_stage=None,**kwargs)[source]#

Bases:_ImageBase

An image with pixels on a regular grid, attached to an Axes.

Parameters:

axAxes

The Axes the image will belong to.

cmapstr orColormap, default:rcParams["image.cmap"] (default:'viridis')

The Colormap instance or registered colormap name used to map scalardata to colors.

normstr orNormalize

Maps luminance to 0-1.

interpolationstr, default:rcParams["image.interpolation"] (default:'auto')

Supported values are 'none', 'auto', 'nearest', 'bilinear','bicubic', 'spline16', 'spline36', 'hanning', 'hamming', 'hermite','kaiser', 'quadric', 'catrom', 'gaussian', 'bessel', 'mitchell','sinc', 'lanczos', 'blackman'.

interpolation_stage{'data', 'rgba'}, default: 'data'

If 'data', interpolationis carried out on the data provided by the user. If 'rgba', theinterpolation is carried out after the colormapping has beenapplied (visual interpolation).

origin{'upper', 'lower'}, default:rcParams["image.origin"] (default:'upper')

Place the [0, 0] index of the array in the upper left or lower leftcorner of the Axes. The convention 'upper' is typically used formatrices and images.

extenttuple, optional

The data axes (left, right, bottom, top) for making image plotsregistered with data plots. Default is to label the pixelcenters with the zero-based row and column indices.

filternormbool, default: True

A parameter for the antigrain image resize filter(see the antigrain documentation).If filternorm is set, the filter normalizes integer values and correctsthe rounding errors. It doesn't do anything with the source floatingpoint values, it corrects only integers according to the rule of 1.0which means that any sum of pixel weights must be equal to 1.0. So,the filter function must produce a graph of the proper shape.

filterradfloat > 0, default: 4

The filter radius for filters that have a radius parameter, i.e. wheninterpolation is one of: 'sinc', 'lanczos' or 'blackman'.

resamplebool, default: False

When True, use a full resampling method. When False, only resample whenthe output image is larger than the input image.

**kwargsArtist properties

Parameters:

colorizercolorizer.Colorizer

get_cursor_data(event)[source]#

Return the image value at the event position orNone if the event isoutside the image.

See also

matplotlib.artist.Artist.get_cursor_data

get_extent()[source]#

Return the image extent as tuple (left, right, bottom, top).

get_window_extent(renderer=None)[source]#

Get the artist's bounding box in display space.

The bounding box's width and height are non-negative.

Subclasses should override for inclusion in the bounding box"tight" calculation. Default is to return an empty boundingbox at 0, 0.

Warning

The extent can change due to any changes in the transform stack, suchas changing the Axes limits, the figure size, the canvas used (as isdone when saving a figure), or the DPI.

Relying on a once-retrieved window extent can lead to unexpectedbehavior in various cases such as interactive figures being resized ormoved to a screen with different dpi, or figures that look fine onscreen render incorrectly when saved to file.

To get accurate results you may need to manually callmatplotlib.figure.Figure.savefig ormatplotlib.figure.Figure.draw_without_rendering to have Matplotlibcompute the rendered size.

make_image(renderer,magnification=1.0,unsampled=False)[source]#

Normalize, rescale, and colormap this image's data for rendering usingrenderer, with the givenmagnification.

Ifunsampled is True, the image will not be scaled, but anappropriate affine transformation will be returned instead.

Returns:

image(M, N, 4)numpy.uint8 array

The RGBA image, resampled unlessunsampled is True.

x, yfloat

The upper left corner where the image should be drawn, in pixelspace.

transAffine2D

The affine transformation from image to pixel space.

set(*,agg_filter=<UNSET>,alpha=<UNSET>,animated=<UNSET>,array=<UNSET>,clim=<UNSET>,clip_box=<UNSET>,clip_on=<UNSET>,clip_path=<UNSET>,cmap=<UNSET>,data=<UNSET>,extent=<UNSET>,filternorm=<UNSET>,filterrad=<UNSET>,gid=<UNSET>,in_layout=<UNSET>,interpolation=<UNSET>,interpolation_stage=<UNSET>,label=<UNSET>,mouseover=<UNSET>,norm=<UNSET>,path_effects=<UNSET>,picker=<UNSET>,rasterized=<UNSET>,resample=<UNSET>,sketch_params=<UNSET>,snap=<UNSET>,transform=<UNSET>,url=<UNSET>,visible=<UNSET>,zorder=<UNSET>)[source]#

Set multiple properties at once.

Supported properties are

Property	Description
agg_filter	a filter function, which takes a (m, n, 3) float array and a dpi value, and returns a (m, n, 3) array and two offsets from the bottom left corner of the image
alpha	float or 2D array-like or None
animated	bool
array	array-like
clim	(vmin: float, vmax: float)
clip_box	BboxBase or None
clip_on	bool
clip_path	Patch or (Path, Transform) or None
cmap	Colormap or str or None
data	array-like orPIL.Image.Image
extent	4-tuple of float
figure	Figure orSubFigure
filternorm	bool
filterrad	positive float
gid	str
in_layout	bool
interpolation	{'auto', 'nearest', 'bilinear', 'bicubic', 'spline16', 'spline36', 'hanning', 'hamming', 'hermite', 'kaiser', 'quadric', 'catrom', 'gaussian', 'bessel', 'mitchell', 'sinc', 'lanczos', 'none'} or None
interpolation_stage	{'data', 'rgba', 'auto'}, default:rcParams["image.interpolation_stage"] (default:'auto')
label	object
mouseover	bool
norm	Normalize or str or None
path_effects	list ofAbstractPathEffect
picker	None or bool or float or callable
rasterized	bool
resample	bool, default:rcParams["image.resample"] (default:True)
sketch_params	(scale: float, length: float, randomness: float)
snap	bool or None
transform	Transform
url	str
visible	bool
zorder	float

set_extent(extent,**kwargs)[source]#

Set the image extent.

Parameters:

extent4-tuple of float

The position and size of the image as tuple(left,right,bottom,top) in data coordinates.

**kwargs

Other parameters from which unit info (i.e., thexunits,yunits,zunits (for 3D Axes),runits andthetaunits (forpolar Axes) entries are applied, if present.

Notes

This updatesAxes.dataLim, and, if autoscaling, setsAxes.viewLimto tightly fit the image, regardless ofdataLim. Autoscalingstate is not changed, so a subsequent call toAxes.autoscale_viewwill redo the autoscaling in accord withdataLim.

classmatplotlib.image.BboxImage(bbox,*,cmap=None,norm=None,colorizer=None,interpolation=None,origin=None,filternorm=True,filterrad=4.0,resample=False,**kwargs)[source]#

Bases:_ImageBase

The Image class whose size is determined by the given bbox.

Parameters:

bboxBboxBase or Callable[RendererBase, BboxBase]

The bbox or a function to generate the bbox

Warning

If usingmatplotlib.artist.Artist.get_window_extent as thecallable ensure that the other artist is drawn first (lower zorder)or you may need to renderer the figure twice to ensure that thecomputed bbox is accurate.

cmapstr orColormap, default:rcParams["image.cmap"] (default:'viridis')

The Colormap instance or registered colormap name used to map scalardata to colors.

normstr orNormalize

Maps luminance to 0-1.

interpolationstr, default:rcParams["image.interpolation"] (default:'auto')

Supported values are 'none', 'auto', 'nearest', 'bilinear','bicubic', 'spline16', 'spline36', 'hanning', 'hamming', 'hermite','kaiser', 'quadric', 'catrom', 'gaussian', 'bessel', 'mitchell','sinc', 'lanczos', 'blackman'.

origin{'upper', 'lower'}, default:rcParams["image.origin"] (default:'upper')

Place the [0, 0] index of the array in the upper left or lower leftcorner of the Axes. The convention 'upper' is typically used formatrices and images.

filternormbool, default: True

A parameter for the antigrain image resize filter(see the antigrain documentation).If filternorm is set, the filter normalizes integer values and correctsthe rounding errors. It doesn't do anything with the source floatingpoint values, it corrects only integers according to the rule of 1.0which means that any sum of pixel weights must be equal to 1.0. So,the filter function must produce a graph of the proper shape.

filterradfloat > 0, default: 4

The filter radius for filters that have a radius parameter, i.e. wheninterpolation is one of: 'sinc', 'lanczos' or 'blackman'.

resamplebool, default: False

When True, use a full resampling method. When False, only resample whenthe output image is larger than the input image.

**kwargsArtist properties

Parameters:

colorizercolorizer.Colorizer

contains(mouseevent)[source]#

Test whether the mouse event occurred within the image.

get_window_extent(renderer=None)[source]#

Get the artist's bounding box in display space.

The bounding box's width and height are non-negative.

Subclasses should override for inclusion in the bounding box"tight" calculation. Default is to return an empty boundingbox at 0, 0.

Warning

The extent can change due to any changes in the transform stack, suchas changing the Axes limits, the figure size, the canvas used (as isdone when saving a figure), or the DPI.

Relying on a once-retrieved window extent can lead to unexpectedbehavior in various cases such as interactive figures being resized ormoved to a screen with different dpi, or figures that look fine onscreen render incorrectly when saved to file.

To get accurate results you may need to manually callmatplotlib.figure.Figure.savefig ormatplotlib.figure.Figure.draw_without_rendering to have Matplotlibcompute the rendered size.

make_image(renderer,magnification=1.0,unsampled=False)[source]#

Normalize, rescale, and colormap this image's data for rendering usingrenderer, with the givenmagnification.

Ifunsampled is True, the image will not be scaled, but anappropriate affine transformation will be returned instead.

Returns:

image(M, N, 4)numpy.uint8 array

The RGBA image, resampled unlessunsampled is True.

x, yfloat

The upper left corner where the image should be drawn, in pixelspace.

transAffine2D

The affine transformation from image to pixel space.

set(*,agg_filter=<UNSET>,alpha=<UNSET>,animated=<UNSET>,array=<UNSET>,clim=<UNSET>,clip_box=<UNSET>,clip_on=<UNSET>,clip_path=<UNSET>,cmap=<UNSET>,data=<UNSET>,filternorm=<UNSET>,filterrad=<UNSET>,gid=<UNSET>,in_layout=<UNSET>,interpolation=<UNSET>,interpolation_stage=<UNSET>,label=<UNSET>,mouseover=<UNSET>,norm=<UNSET>,path_effects=<UNSET>,picker=<UNSET>,rasterized=<UNSET>,resample=<UNSET>,sketch_params=<UNSET>,snap=<UNSET>,transform=<UNSET>,url=<UNSET>,visible=<UNSET>,zorder=<UNSET>)[source]#

Set multiple properties at once.

Supported properties are

Property	Description
agg_filter	a filter function, which takes a (m, n, 3) float array and a dpi value, and returns a (m, n, 3) array and two offsets from the bottom left corner of the image
alpha	float or 2D array-like or None
animated	bool
array	array-like
clim	(vmin: float, vmax: float)
clip_box	BboxBase or None
clip_on	bool
clip_path	Patch or (Path, Transform) or None
cmap	Colormap or str or None
data	array-like orPIL.Image.Image
figure	Figure orSubFigure
filternorm	bool
filterrad	positive float
gid	str
in_layout	bool
interpolation	{'auto', 'nearest', 'bilinear', 'bicubic', 'spline16', 'spline36', 'hanning', 'hamming', 'hermite', 'kaiser', 'quadric', 'catrom', 'gaussian', 'bessel', 'mitchell', 'sinc', 'lanczos', 'none'} or None
interpolation_stage	{'data', 'rgba', 'auto'}, default:rcParams["image.interpolation_stage"] (default:'auto')
label	object
mouseover	bool
norm	Normalize or str or None
path_effects	list ofAbstractPathEffect
picker	None or bool or float or callable
rasterized	bool
resample	bool, default:rcParams["image.resample"] (default:True)
sketch_params	(scale: float, length: float, randomness: float)
snap	bool or None
transform	Transform
url	str
visible	bool
zorder	float

classmatplotlib.image.FigureImage(fig,*,cmap=None,norm=None,colorizer=None,offsetx=0,offsety=0,origin=None,**kwargs)[source]#

Bases:_ImageBase

An image attached to a figure.

cmap is a colors.Colormap instancenorm is a colors.Normalize instance to map luminance to 0-1

kwargs are an optional list of Artist keyword args

get_extent()[source]#

Return the image extent as tuple (left, right, bottom, top).

make_image(renderer,magnification=1.0,unsampled=False)[source]#

Normalize, rescale, and colormap this image's data for rendering usingrenderer, with the givenmagnification.

Ifunsampled is True, the image will not be scaled, but anappropriate affine transformation will be returned instead.

Returns:

image(M, N, 4)numpy.uint8 array

The RGBA image, resampled unlessunsampled is True.

x, yfloat

The upper left corner where the image should be drawn, in pixelspace.

transAffine2D

The affine transformation from image to pixel space.

set(*,agg_filter=<UNSET>,alpha=<UNSET>,animated=<UNSET>,array=<UNSET>,clim=<UNSET>,clip_box=<UNSET>,clip_on=<UNSET>,clip_path=<UNSET>,cmap=<UNSET>,data=<UNSET>,filternorm=<UNSET>,filterrad=<UNSET>,gid=<UNSET>,in_layout=<UNSET>,interpolation=<UNSET>,interpolation_stage=<UNSET>,label=<UNSET>,mouseover=<UNSET>,norm=<UNSET>,path_effects=<UNSET>,picker=<UNSET>,rasterized=<UNSET>,resample=<UNSET>,sketch_params=<UNSET>,snap=<UNSET>,transform=<UNSET>,url=<UNSET>,visible=<UNSET>,zorder=<UNSET>)[source]#

Set multiple properties at once.

Supported properties are

Property	Description
agg_filter	a filter function, which takes a (m, n, 3) float array and a dpi value, and returns a (m, n, 3) array and two offsets from the bottom left corner of the image
alpha	float or 2D array-like or None
animated	bool
array	array-like
clim	(vmin: float, vmax: float)
clip_box	BboxBase or None
clip_on	bool
clip_path	Patch or (Path, Transform) or None
cmap	Colormap or str or None
data	unknown
figure	Figure orSubFigure
filternorm	bool
filterrad	positive float
gid	str
in_layout	bool
interpolation	{'auto', 'nearest', 'bilinear', 'bicubic', 'spline16', 'spline36', 'hanning', 'hamming', 'hermite', 'kaiser', 'quadric', 'catrom', 'gaussian', 'bessel', 'mitchell', 'sinc', 'lanczos', 'none'} or None
interpolation_stage	{'data', 'rgba', 'auto'}, default:rcParams["image.interpolation_stage"] (default:'auto')
label	object
mouseover	bool
norm	Normalize or str or None
path_effects	list ofAbstractPathEffect
picker	None or bool or float or callable
rasterized	bool
resample	bool, default:rcParams["image.resample"] (default:True)
sketch_params	(scale: float, length: float, randomness: float)
snap	bool or None
transform	Transform
url	str
visible	bool
zorder	float

set_data(A)[source]#

Set the image array.

zorder=0#

classmatplotlib.image.NonUniformImage(ax,*,interpolation='nearest',**kwargs)[source]#

Bases:AxesImage

An image with pixels on a rectilinear grid.

In contrast toAxesImage, where pixels are on a regular grid,NonUniformImage allows rows and columns with individual heights / widths.

See alsoImage nonuniform.

Parameters:

axAxes

The Axes the image will belong to.

interpolation{'nearest', 'bilinear'}, default: 'nearest'

The interpolation scheme used in the resampling.

**kwargs

All other keyword arguments are identical to those ofAxesImage.

get_cursor_data(event)[source]#

Return the image value at the event position orNone if the event isoutside the image.

See also

matplotlib.artist.Artist.get_cursor_data

get_extent()[source]#

Return the image extent as tuple (left, right, bottom, top).

make_image(renderer,magnification=1.0,unsampled=False)[source]#

Normalize, rescale, and colormap this image's data for rendering usingrenderer, with the givenmagnification.

Ifunsampled is True, the image will not be scaled, but anappropriate affine transformation will be returned instead.

Returns:

image(M, N, 4)numpy.uint8 array

The RGBA image, resampled unlessunsampled is True.

x, yfloat

The upper left corner where the image should be drawn, in pixelspace.

transAffine2D

The affine transformation from image to pixel space.

set(*,agg_filter=<UNSET>,alpha=<UNSET>,animated=<UNSET>,array=<UNSET>,clim=<UNSET>,clip_box=<UNSET>,clip_on=<UNSET>,clip_path=<UNSET>,cmap=<UNSET>,data=<UNSET>,extent=<UNSET>,filternorm=<UNSET>,filterrad=<UNSET>,gid=<UNSET>,in_layout=<UNSET>,interpolation=<UNSET>,interpolation_stage=<UNSET>,label=<UNSET>,mouseover=<UNSET>,norm=<UNSET>,path_effects=<UNSET>,picker=<UNSET>,rasterized=<UNSET>,resample=<UNSET>,sketch_params=<UNSET>,snap=<UNSET>,transform=<UNSET>,url=<UNSET>,visible=<UNSET>,zorder=<UNSET>)[source]#

Set multiple properties at once.

Supported properties are

Property	Description
agg_filter	a filter function, which takes a (m, n, 3) float array and a dpi value, and returns a (m, n, 3) array and two offsets from the bottom left corner of the image
alpha	float or 2D array-like or None
animated	bool
array	unknown
clim	(vmin: float, vmax: float)
clip_box	BboxBase or None
clip_on	bool
clip_path	Patch or (Path, Transform) or None
cmap	unknown
data	unknown
extent	4-tuple of float
figure	Figure orSubFigure
filternorm	unknown
filterrad	unknown
gid	str
in_layout	bool
interpolation	{'nearest', 'bilinear'} or None
interpolation_stage	{'data', 'rgba', 'auto'}, default:rcParams["image.interpolation_stage"] (default:'auto')
label	object
mouseover	bool
norm	unknown
path_effects	list ofAbstractPathEffect
picker	None or bool or float or callable
rasterized	bool
resample	bool, default:rcParams["image.resample"] (default:True)
sketch_params	(scale: float, length: float, randomness: float)
snap	bool or None
transform	Transform
url	str
visible	bool
zorder	float

set_array(*args)[source]#

Retained for backwards compatibility - use set_data instead.

Parameters:

Aarray-like

set_cmap(cmap)[source]#

Set the colormap for luminance data.

Parameters:

cmapColormap or str or None

set_data(x,y,A)[source]#

Set the grid for the pixel centers, and the pixel values.

Parameters:

x, y1D array-like

Monotonic arrays of shapes (N,) and (M,), respectively, specifyingpixel centers.

Aarray-like

(M, N)ndarray or masked array of values to becolormapped, or (M, N, 3) RGB array, or (M, N, 4) RGBA array.

set_filternorm(filternorm)[source]#

Set whether the resize filter normalizes the weights.

See help forimshow.

Parameters:

filternormbool

set_filterrad(filterrad)[source]#

Set the resize filter radius (only applicable to someinterpolation schemes).

See help forimshow.

Parameters:

filterradpositive float

set_interpolation(s)[source]#

Parameters:

s{'nearest', 'bilinear'} or None

If None, usercParams["image.interpolation"] (default:'auto').

set_norm(norm)[source]#

Set the normalization instance.

Parameters:

normNormalize or str or None

Notes

If there are any colorbars using the mappable for this norm, settingthe norm of the mappable will reset the norm, locator, and formatterson the colorbar to default.

classmatplotlib.image.PcolorImage(ax,x=None,y=None,A=None,*,cmap=None,norm=None,colorizer=None,**kwargs)[source]#

Bases:AxesImage

Make a pcolor-style plot with an irregular rectangular grid.

This uses a variation of the original irregular image code,and it is used by pcolorfast for the corresponding grid type.

Parameters:

axAxes

The Axes the image will belong to.

x, y1D array-like, optional

Monotonic arrays of length N+1 and M+1, respectively, specifyingrectangle boundaries. If not given, will default torange(N+1) andrange(M+1), respectively.

Aarray-like

The data to be color-coded. The interpretation depends on theshape:

• (M, N)ndarray or masked array: values to be colormapped

• (M, N, 3): RGB array

• (M, N, 4): RGBA array

cmapstr orColormap, default:rcParams["image.cmap"] (default:'viridis')

The Colormap instance or registered colormap name used to mapscalar data to colors.

normstr orNormalize

Maps luminance to 0-1.

**kwargsArtist properties

get_cursor_data(event)[source]#

Return the image value at the event position orNone if the event isoutside the image.

See also

matplotlib.artist.Artist.get_cursor_data

make_image(renderer,magnification=1.0,unsampled=False)[source]#

Normalize, rescale, and colormap this image's data for rendering usingrenderer, with the givenmagnification.

Ifunsampled is True, the image will not be scaled, but anappropriate affine transformation will be returned instead.

Returns:

image(M, N, 4)numpy.uint8 array

The RGBA image, resampled unlessunsampled is True.

x, yfloat

The upper left corner where the image should be drawn, in pixelspace.

transAffine2D

The affine transformation from image to pixel space.

set(*,agg_filter=<UNSET>,alpha=<UNSET>,animated=<UNSET>,array=<UNSET>,clim=<UNSET>,clip_box=<UNSET>,clip_on=<UNSET>,clip_path=<UNSET>,cmap=<UNSET>,data=<UNSET>,extent=<UNSET>,filternorm=<UNSET>,filterrad=<UNSET>,gid=<UNSET>,in_layout=<UNSET>,interpolation=<UNSET>,interpolation_stage=<UNSET>,label=<UNSET>,mouseover=<UNSET>,norm=<UNSET>,path_effects=<UNSET>,picker=<UNSET>,rasterized=<UNSET>,resample=<UNSET>,sketch_params=<UNSET>,snap=<UNSET>,transform=<UNSET>,url=<UNSET>,visible=<UNSET>,zorder=<UNSET>)[source]#

Set multiple properties at once.

Supported properties are

Property	Description
agg_filter	a filter function, which takes a (m, n, 3) float array and a dpi value, and returns a (m, n, 3) array and two offsets from the bottom left corner of the image
alpha	float or 2D array-like or None
animated	bool
array	unknown
clim	(vmin: float, vmax: float)
clip_box	BboxBase or None
clip_on	bool
clip_path	Patch or (Path, Transform) or None
cmap	Colormap or str or None
data	unknown
extent	4-tuple of float
figure	Figure orSubFigure
filternorm	bool
filterrad	positive float
gid	str
in_layout	bool
interpolation	{'auto', 'nearest', 'bilinear', 'bicubic', 'spline16', 'spline36', 'hanning', 'hamming', 'hermite', 'kaiser', 'quadric', 'catrom', 'gaussian', 'bessel', 'mitchell', 'sinc', 'lanczos', 'none'} or None
interpolation_stage	{'data', 'rgba', 'auto'}, default:rcParams["image.interpolation_stage"] (default:'auto')
label	object
mouseover	bool
norm	Normalize or str or None
path_effects	list ofAbstractPathEffect
picker	None or bool or float or callable
rasterized	bool
resample	bool, default:rcParams["image.resample"] (default:True)
sketch_params	(scale: float, length: float, randomness: float)
snap	bool or None
transform	Transform
url	str
visible	bool
zorder	float

set_array(*args)[source]#

Retained for backwards compatibility - use set_data instead.

Parameters:

Aarray-like

set_data(x,y,A)[source]#

Set the grid for the rectangle boundaries, and the data values.

Parameters:

x, y1D array-like, optional

Monotonic arrays of length N+1 and M+1, respectively, specifyingrectangle boundaries. If not given, will default torange(N+1) andrange(M+1), respectively.

Aarray-like

The data to be color-coded. The interpretation depends on theshape:

• (M, N)ndarray or masked array: values to be colormapped

• (M, N, 3): RGB array

• (M, N, 4): RGBA array

matplotlib.image.composite_images(images,renderer,magnification=1.0)[source]#

Composite a number of RGBA images into one. The images arecomposited in the order in which they appear in theimages list.

Parameters:

imageslist of Images

Each must have amake_image method. For each image,can_composite should returnTrue, though this is notenforced by this function. Each image must have a purelyaffine transformation with no shear.

rendererRendererBase

magnificationfloat, default: 1

The additional magnification to apply for the renderer in use.

Returns:

image(M, N, 4)numpy.uint8 array

The composited RGBA image.

offset_x, offset_yfloat

The (left, bottom) offset where the composited image should be placedin the output figure.

matplotlib.image.imread(fname,format=None)[source]#

Read an image from a file into an array.

Note

This function exists for historical reasons. It is recommended tousePIL.Image.open instead for loading images.

Parameters:

fnamestr or file-like

The image file to read: a filename, a URL or a file-like object openedin read-binary mode.

Passing a URL is deprecated. Please open the URLfor reading and pass the result to Pillow, e.g. withnp.array(PIL.Image.open(urllib.request.urlopen(url))).

formatstr, optional

The image file format assumed for reading the data. The image isloaded as a PNG file ifformat is set to "png", iffname is a pathor opened file with a ".png" extension, or if it is a URL. In allother cases,format is ignored and the format is auto-detected byPIL.Image.open.

Returns:

numpy.array

The image data. The returned array has shape

• (M, N) for grayscale images.

• (M, N, 3) for RGB images.

• (M, N, 4) for RGBA images.

PNG images are returned as float arrays (0-1). All other formats arereturned as int arrays, with a bit depth determined by the file'scontents.

matplotlib.image.imsave(fname,arr,vmin=None,vmax=None,cmap=None,format=None,origin=None,dpi=100,*,metadata=None,pil_kwargs=None)[source]#

Colormap and save an array as an image file.

RGB(A) images are passed through. Single channel images will becolormapped according tocmap andnorm.

Note

If you want to save a single channel image as gray scale please use animage I/O library (such as pillow, tifffile, or imageio) directly.

Parameters:

fnamestr or path-like or file-like

A path or a file-like object to store the image in.Ifformat is not set, then the output format is inferred from theextension offname, if any, and fromrcParams["savefig.format"] (default:'png') otherwise.Ifformat is set, it determines the output format.

arrarray-like

The image data. Accepts NumPy arrays or sequences(e.g., lists or tuples). The shape can be one ofMxN (luminance), MxNx3 (RGB) or MxNx4 (RGBA).

vmin, vmaxfloat, optional

vmin andvmax set the color scaling for the image by fixing thevalues that map to the colormap color limits. If eithervminorvmax is None, that limit is determined from thearrmin/max value.

cmapstr orColormap, default:rcParams["image.cmap"] (default:'viridis')

A Colormap instance or registered colormap name. The colormapmaps scalar data to colors. It is ignored for RGB(A) data.

formatstr, optional

The file format, e.g. 'png', 'pdf', 'svg', ... The behavior when thisis unset is documented underfname.

origin{'upper', 'lower'}, default:rcParams["image.origin"] (default:'upper')

Indicates whether the(0,0) index of the array is in the upperleft or lower left corner of the Axes.

dpifloat

The DPI to store in the metadata of the file. This does not affect theresolution of the output image. Depending on file format, this may berounded to the nearest integer.

metadatadict, optional

Metadata in the image file. The supported keys depend on the outputformat, see the documentation of the respective backends for moreinformation.Currently only supported for "png", "pdf", "ps", "eps", and "svg".

pil_kwargsdict, optional

Keyword arguments passed toPIL.Image.Image.save. If the 'pnginfo'key is present, it completely overridesmetadata, including thedefault 'Software' key.

matplotlib.image.pil_to_array(pilImage)[source]#

Load aPIL image and return it as a numpy int array.

Returns:

numpy.array

The array shape depends on the image type:

• (M, N) for grayscale images.

• (M, N, 3) for RGB images.

• (M, N, 4) for RGBA images.

matplotlib.image.thumbnail(infile,thumbfile,scale=0.1,interpolation='bilinear',preview=False)[source]#

Make a thumbnail of image ininfile with output filenamethumbfile.

SeeImage thumbnail.

Parameters:

infilestr or file-like

The image file. Matplotlib relies onPillow for image reading, andthus supports a wide range of file formats, including PNG, JPG, TIFFand others.

thumbfilestr or file-like

The thumbnail filename.

scalefloat, default: 0.1

The scale factor for the thumbnail.

interpolationstr, default: 'bilinear'

The interpolation scheme used in the resampling. See theinterpolation parameter ofimshow for possible values.

previewbool, default: False

If True, the default backend (presumably a user interfacebackend) will be used which will cause a figure to be raised ifshow is called. If it is False, the figure iscreated usingFigureCanvasBase and the drawing backend is selectedasFigure.savefig would normally do.

Returns:

Figure

The figure instance containing the thumbnail.