Generally, the geometric feature parameters of the target area include perimeter, area, longest axis, azimuth, boundary matrix, and shape coefficient.
-The length of the outer boundary of the target area of the perimeter. It can be expressed by the sum of the distance between two adjacent pixels on the outer boundary.
-The area can be expressed by the number of pixels on the platform in the target region. After the image is tagged, the area parameter can be easily extracted using the embedding method.
-The maximum extension length of the target region of the longest axis, that is, the line between the two image points with the largest distance in the outer world of the target region. If the target region's outer boundary is known, the maximum outer boundary point can be obtained through the repeated comparison of the distance between different outer boundary pixels. Therefore, the off pound extracted by the longest axis parameter is the extraction of the outer boundary of the target region.
-The angle between the maximum axis and the X axis of the target region of the azimuth angle.
-The minimum matrix of the target region enclosed by the boundary matrix. It intuitively reflects the flat degree of the target area. Obtain the two outer boundary tangent parallel to the longest axis and the two outer boundary tangent perpendicular to the longest axis. The intersection of the four lines forms the boundary matrix.
-The shape coefficient is used to reflect the degree of stenosis of the target. The ratio of the Area parameter to the square of the perimeter parameter.
From: http://blog.csdn.net/zernike/article/details/1207540