三类

basic data types
- 通常是已经从c++中继承的
- int, float, etc
- vectors,matrices,simple geometric concepts like points, rectangles,sizes
helper objects
- 代表一些更抽象的概念
- garbage-collecting pointer class
- range objects used for slicing
- abstractions such as termination criteria
large array types
- contain arrays or other assemblies of primitives or, more often, the basic data types
cv::Mat class
- 表示任意维数组
- 包含其他对象,如稀疏矩阵 cv::SparseMat class
模板类 cv::Vec<>
- 代表 fixed vector classes
- 用于一些编译时就已知维度的小向量
fixed matrix classes
- associated with the template cv::Matx<>.
- certain specific small matrix operations
- the dimensionality of the fixed matrix classes must be known at compile time
point classes
- their members are accessed by named variables (mypoint.x, mypoint.y, etc.) rather than by a vector index (myvec[0],myvec[1], etc.)
cv::Scalar
- four-dimensional point
- generate an arbitrary four-component vector
cv::Size and cv::Rect
- cv::Size -- having data members width and height rather than x and y
- cv::Rect -- all four
- cv::RotatedRect not axis-aligned 非抽对称

point classes

可以是任意类型的
有两个模板
- 2-D
- 3-D
- cv::Point2i or cv::Point3f (第一个数-维度, 第二个数-精度)
- b is an unsigned character, s is a short integer, i is a 32-bit integer, f is a 32-bit floating-point number, and d is a 64- bit floating-point number

Example

Operation	Example
Default constructors	cv::Point2i p; cv::Point3f p;
Copy constructor	cv::Point3f p2( p1 );
Value constructors	cv::Point2i p( x0, x1 ); cv::Point3d p( x0, x1, x2 );
Cast to the fixed vector classes	(cv::Vec3f) p;
Member access	p.x; p.y; // and for three-dimensional // point classes: p.z
Dot product	float x = p1.dot( p2 )
Double-precision dot product	double x = p1.ddot( p2 )
Cross product	p1.cross( p2 ) // (for three-dimensional point // classes only)
Query if point p is inside rectangle r	p.inside( r ) // (for two-dimensional point // classes only)