图像配准

使用强度相关性、特征匹配或控制点映射来对齐两个图像

Image Processing Toolbox™ 和 Computer Vision Toolbox™ 一共提供四种图像配准解决方案：使用图像配准器的交互式配准、基于强度的自动图像配准、控制点配准和自动特征匹配。有关选择使用哪种方法的帮助，请参阅Approaches to Registering Images。

App

图像配准器

配准二维灰度图像

函数

全部展开

自动配准

`imregister`	Intensity-based image registration
`imregconfig`	Configurations for intensity-based registration
`imregtform`	Estimate geometric transformation that aligns two 2-D or 3-D images
`imregcorr`	Estimate geometric transformation that aligns two 2-D images using phase correlation
`imregdemons`	Estimate displacement field that aligns two 2-D or 3-D images
`imregmtb`	Register 2-D images using median threshold bitmaps
`normxcorr2`	归一化二维互相关性
`MattesMutualInformation`	Mattes mutual information metric configuration
`MeanSquares`	Mean square error metric configuration
`RegularStepGradientDescent`	Regular step gradient descent optimizer configuration
`OnePlusOneEvolutionary`	One-plus-one evolutionary optimizer configuration

控制点配准

`cpselect`	Control Point Selection tool
`fitgeotrans`	对控制点对组进行几何变换拟合
`cpcorr`	Tune control point locations using cross-correlation
`cpstruct2pairs`	Extract valid control point pairs from `cpstruct` structure
`imwarp`	对图像应用几何变换

空间参照

`imref2d`	Reference 2-D image to world coordinates
`imref3d`	Reference 3-D image to world coordinates

主题

以交互方式配准图像

Register Images Using Registration Estimator App

Align a pair of images using feature-based registration techniques in the Registration Estimator app.
Techniques Supported by Registration Estimator
The Registration Estimator app provides algorithms for feature-based, intensity-based, and nonrigid registration.

使用基于强度的优化配准图像

Intensity-Based Automatic Image Registration
Intensity-based automatic image registration uses a similarity metric, an optimizer, and a transformation type to register two images iteratively.
为基于强度的图像配准创建优化器和指标
选择适用于单模态或多模态图像的图像指标和优化器。
Use Phase Correlation as Preprocessing Step in Registration
Use phase correlation to estimate an initial transformation when images are severely misaligned.
Registering an Image Using Normalized Cross-Correlation
Determine the translation needed to align a cropped subset of an image with the larger image.

使用控制点映射配准图像

Control Point Registration
To determine the parameters of a geometric transformation, you can pick corresponding pairs of points in two images.
Geometric Transformation Types for Control Point Registration
Control point registration can infer the parameters for similarity, affine, projective, polynomial, piecewise linear, and local weighted mean transformations.
控制点选择过程

要以交互方式指定一对图像中的控制点，请使用控制点选择工具。
Use Cross-Correlation to Improve Control Point Placement
Fine-tune your control point selections using cross-correlation.

精选示例

Find Image Rotation and Scale

Use control points to align two images that differ by a rotation and scale change.

打开实时脚本

Register Multimodal MRI Images

Align two multimodal MRI images to a common coordinate system using automatic intensity-based image registration.

打开实时脚本

Register Multimodal 3-D Medical Images

Align two volumetric images using automatic intensity-based image registration.

打开实时脚本

Register Images with Projection Distortion Using Control Points

Register two images by selecting control points common to both images and inferring a geometric transformation that aligns the control points.

打开实时脚本

Find Image Rotation and Scale Using Automated Feature Matching

Automatically determine the geometric transformation between a pair of images. When one image is distorted relative to another by rotation and scale, use detectSURFFeatures and estgeotform2d to find the rotation angle and scale factor. You can then transform the distorted image to recover the original image.

Feature Based Panoramic Image Stitching

Automatically create a panorama using feature based image registration techniques.

Video Stabilization Using Point Feature Matching

Stabilize a video that was captured from a jittery platform. One way to stabilize a video is to track a salient feature in the image and use this as an anchor point to cancel out all perturbations relative to it. This procedure, however, must be bootstrapped with knowledge of where such a salient feature lies in the first video frame. In this example, we explore a method of video stabilization that works without any such a priori knowledge. It instead automatically searches for the "background plane" in a video sequence, and uses its observed distortion to correct for camera motion.