opencv之cvtColor()函数

概述

opencv使用cvtColor()函数对图像进行色彩空间的转换。

API说明

C++ API如下表所示：

void cv::cvtColor	(	InputArray	src,	（原始图像，格式：8-bit unsigned / 16-bit unsigned / single-precision floating-point）
		OutputArray	dst,	（目标图像：size与depth和原始图像要相等）
		int	code,	（色彩空间转换码）（ColorConversionCodes）
		int	dstCn=0	（目标图像的通道数，此参数为0时，目标图像的通道数由原始图像自动生成）
	)

python API如下表所示：

dst	=	cv.cvtColor(	src, code[, dst[, dstCn]]	)

功能说明

• 将输入图像从一个色彩空间转换到另一个色彩空间。
• 在处理RGB图像时要注意色彩通道的顺序(RGB或BGR)。
• opencv中的默认色彩空间实际是BGR，但是常常以RGB提及。
• 对于一个24位标准彩色图像而言，1个像素对应3个字节（8位），每个字节依次对应B，G和R分量。

• R，G，和B分量的取值范围一般如下：

  • 对于CV_8U 图像：0 到 255
  • 对于CV_16U 图像：0 到 65535
  • 对于CV_32F 图像：0 到 1
  • 对于线性变换，该取值范围 不重要。
  • 对于非线性变换，输入的rgb图像应当先归一化到正确的取值范围，从而得到正确的结果。
  • 例如：RGB → L*u*v*，输入图像：8-bit image，目标图像：32-bit floating-point image，在调用 cvtColor()函数之前，应当将输入图像的分量的取值范围从0~255缩到0~1：
    img *= 1./255;
    cvtColor(img, img, COLOR_BGR2Luv);

• 如果对8-bit images使用cvtColor()函数，转换会丢失部分信息，一般不明显。推荐先将图像转换成32-bit images，再进行色彩空间转换，最后再把图像转换成8-bit images。

• 如果转换后添加了透明通道，其值会被设置为对应通道取值范围的最大值（255 for CV_8U, 65535 for CV_16U, 1 for CV_32F）。

官网说明

Color conversions

官网样例

camshiftdemo.cpp
ffilldemo.cpp, hog.cpp
houghcircles.cpp
houghlines.cpp
lkdemo.cpp
Sobel_Demo.cpp
train_HOG.cpp
watershed.cpp.

附录一：色彩空间转换代码表

转换代码	转换说明
COLOR_BGR2BGRA	add alpha channel to RGB or BGR image
COLOR_RGB2RGBA
COLOR_BGRA2BGR	remove alpha channel from RGB or BGR image
COLOR_RGBA2RGB
COLOR_BGR2RGBA	convert between RGB and BGR colorspaces(with or without alpha channel)
COLOR_RGB2BGRA
COLOR_RGBA2BGR
COLOR_BGRA2RGB
COLOR_BGR2RGB
COLOR_RGB2BGR
COLOR_BGRA2RGBA
COLOR_RGBA2BGRA
COLOR_BGR2GRAY	convert between RGB/BGR and grayscale,官网代码样例
COLOR_RGB2GRAY
COLOR_GRAY2BGR
COLOR_GRAY2RGB
COLOR_GRAY2BGRA
COLOR_GRAY2RGBA
COLOR_BGRA2GRAY
COLOR_RGBA2GRAY
COLOR_BGR2BGR565	convert between RGB/BGR and BGR565(16-bit images)
COLOR_RGB2BGR565
COLOR_BGR5652BGR
COLOR_BGR5652RGB
COLOR_BGRA2BGR565
COLOR_RGBA2BGR565
COLOR_BGR5652BGRA
COLOR_BGR5652RGBA
COLOR_GRAY2BGR565	convert between grayscale and BGR565(16-bit images)
COLOR_BGR5652GRAY
COLOR_BGR2BGR555	convert between RGB/BGR and BGR555(16-bit images)
COLOR_RGB2BGR555
COLOR_BGR5552BGR
COLOR_BGR5552RGB
COLOR_BGRA2BGR555
COLOR_RGBA2BGR555
COLOR_BGR5552BGRA
COLOR_BGR5552RGBA
COLOR_GRAY2BGR555	convert between grayscale and BGR555(16-bit images)
COLOR_BGR5552GRAY
COLOR_BGR2XYZ	convert RGB/BGR to CIEXYZ,官网代码样例
COLOR_RGB2XYZ
COLOR_XYZ2BGR
COLOR_XYZ2RGB
COLOR_BGR2YCrCb	convert RGB/BGR to luma-chroma(akaYCC),官网代码样例
COLOR_RGB2YCrCb
COLOR_YCrCb2BGR
COLOR_YCrCb2RGB
COLOR_BGR2HSV	convert RGB/BGR to HSV(hue saturation value),官网代码样例
COLOR_RGB2HSV
COLOR_BGR2Lab	convert RGB/BGR to CIELab,官网代码样例
COLOR_RGB2Lab
COLOR_BGR2Luv	convert RGB/BGR to CIELuv,官网代码样例
COLOR_RGB2Luv
COLOR_BGR2HLS	convert RGB/BGR to HLS(hue lightness saturation),官网代码样例
COLOR_RGB2HLS
COLOR_HSV2BGR	backward conversions to RGB/BGR
COLOR_HSV2RGB
COLOR_Lab2BGR
COLOR_Lab2RGB
COLOR_Luv2BGR
COLOR_Luv2RGB
COLOR_HLS2BGR
COLOR_HLS2RGB
COLOR_BGR2HSV_FULL
COLOR_RGB2HSV_FULL
COLOR_BGR2HLS_FULL
COLOR_RGB2HLS_FULL
COLOR_HSV2BGR_FULL
COLOR_HSV2RGB_FULL
COLOR_HLS2BGR_FULL
COLOR_HLS2RGB_FULL
COLOR_LBGR2Lab
COLOR_LRGB2Lab
COLOR_LBGR2Luv
COLOR_LRGB2Luv
COLOR_Lab2LBGR
COLOR_Lab2LRGB
COLOR_Luv2LBGR
COLOR_Luv2LRGB
COLOR_BGR2YUV	convert between RGB/BGR and YUV
COLOR_RGB2YUV
COLOR_YUV2BGR
COLOR_YUV2RGB
COLOR_YUV2RGB_NV12	YUV4:2:0 family to RGB.
COLOR_YUV2BGR_NV12
COLOR_YUV2RGB_NV21
COLOR_YUV2BGR_NV21
COLOR_YUV420sp2RGB
COLOR_YUV420sp2BGR
COLOR_YUV2RGBA_NV12
COLOR_YUV2BGRA_NV12
COLOR_YUV2RGBA_NV21
COLOR_YUV2BGRA_NV21
COLOR_YUV420sp2RGBA
COLOR_YUV420sp2BGRA
COLOR_YUV2RGB_YV12
COLOR_YUV2BGR_YV12
COLOR_YUV2RGB_IYUV
COLOR_YUV2BGR_IYUV
COLOR_YUV2RGB_I420
COLOR_YUV2BGR_I420
COLOR_YUV420p2RGB
COLOR_YUV420p2BGR
COLOR_YUV2RGBA_YV12
COLOR_YUV2BGRA_YV12
COLOR_YUV2RGBA_IYUV
COLOR_YUV2BGRA_IYUV
COLOR_YUV2RGBA_I420
COLOR_YUV2BGRA_I420
COLOR_YUV420p2RGBA
COLOR_YUV420p2BGRA
COLOR_YUV2GRAY_420
COLOR_YUV2GRAY_NV21
COLOR_YUV2GRAY_NV12
COLOR_YUV2GRAY_YV12
COLOR_YUV2GRAY_IYUV
COLOR_YUV2GRAY_I420
COLOR_YUV420sp2GRAY
COLOR_YUV420p2GRAY
COLOR_YUV2RGB_UYVY	YUV4:2:2 family to RGB.
COLOR_YUV2BGR_UYVY
COLOR_YUV2RGB_Y422
COLOR_YUV2BGR_Y422
COLOR_YUV2RGB_UYNV
COLOR_YUV2BGR_UYNV
COLOR_YUV2RGBA_UYVY
COLOR_YUV2BGRA_UYVY
COLOR_YUV2RGBA_Y422
COLOR_YUV2BGRA_Y422
COLOR_YUV2RGBA_UYNV
COLOR_YUV2BGRA_UYNV
COLOR_YUV2RGB_YUY2
COLOR_YUV2BGR_YUY2
COLOR_YUV2RGB_YVYU
COLOR_YUV2BGR_YVYU
COLOR_YUV2RGB_YUYV
COLOR_YUV2BGR_YUYV
COLOR_YUV2RGB_YUNV
COLOR_YUV2BGR_YUNV
COLOR_YUV2RGBA_YUY2
COLOR_YUV2BGRA_YUY2
COLOR_YUV2RGBA_YVYU
COLOR_YUV2BGRA_YVYU
COLOR_YUV2RGBA_YUYV
COLOR_YUV2BGRA_YUYV
COLOR_YUV2RGBA_YUNV
COLOR_YUV2BGRA_YUNV
COLOR_YUV2GRAY_UYVY
COLOR_YUV2GRAY_YUY2
COLOR_YUV2GRAY_Y422
COLOR_YUV2GRAY_UYNV
COLOR_YUV2GRAY_YVYU
COLOR_YUV2GRAY_YUYV
COLOR_YUV2GRAY_YUNV
COLOR_RGBA2mRGBA	alpha premultiplication
COLOR_mRGBA2RGBA
COLOR_RGB2YUV_I420	RGB to YUV4:2:0 family.
COLOR_BGR2YUV_I420
COLOR_RGB2YUV_IYUV
COLOR_BGR2YUV_IYUV
COLOR_RGBA2YUV_I420
COLOR_BGRA2YUV_I420
COLOR_RGBA2YUV_IYUV
COLOR_BGRA2YUV_IYUV
COLOR_RGB2YUV_YV12
COLOR_BGR2YUV_YV12
COLOR_RGBA2YUV_YV12
COLOR_BGRA2YUV_YV12
COLOR_BayerBG2BGR	Demosaicing.
COLOR_BayerGB2BGR
COLOR_BayerRG2BGR
COLOR_BayerGR2BGR
COLOR_BayerBG2RGB
COLOR_BayerGB2RGB
COLOR_BayerRG2RGB
COLOR_BayerGR2RGB
COLOR_BayerBG2GRAY
COLOR_BayerGB2GRAY
COLOR_BayerRG2GRAY
COLOR_BayerGR2GRAY
COLOR_BayerBG2BGR_VNG	Demosaicing using Variable Numberof Gradients.
COLOR_BayerGB2BGR_VNG
COLOR_BayerRG2BGR_VNG
COLOR_BayerGR2BGR_VNG
COLOR_BayerBG2RGB_VNG
COLOR_BayerGB2RGB_VNG
COLOR_BayerRG2RGB_VNG
COLOR_BayerGR2RGB_VNG
COLOR_BayerBG2BGR_EA	Edge-Aware Demosaicing.
COLOR_BayerGB2BGR_EA
COLOR_BayerRG2BGR_EA
COLOR_BayerGR2BGR_EA
COLOR_BayerBG2RGB_EA
COLOR_BayerGB2RGB_EA
COLOR_BayerRG2RGB_EA
COLOR_BayerGR2RGB_EA
COLOR_BayerBG2BGRA	Demosaicing with alpha channel.
COLOR_BayerGB2BGRA
COLOR_BayerRG2BGRA
COLOR_BayerGR2BGRA
COLOR_BayerBG2RGBA
COLOR_BayerGB2RGBA
COLOR_BayerRG2RGBA
COLOR_BayerGR2RGBA
COLOR_COLORCVT_MAX