Nondestructive Detection of Masson Pine Seedlings Morphological Indexes based on 3D-Reconstruction and SVR

China's forest area is wide, and the plantation area ranks first in the world. Masson Pine is one of the widely distributed afforestation varieties of trees in China. Obtaining morphological indexes of Masson Pine seedlings is helpful to select excellent afforestation seedlings and further improve the afforestation effect of seedlings. In order to realize the rapid and accurate evaluation of morphological indexes of Masson Pine seedlings a set of nondestructive detection system for morphological indexes of Masson Pine seedlings was designed, which integrated machine vision technology and machine learning technology. Firstly, an image acquisition hardware experimental platform is established and the camera is calibrated; then the software system performs preprocessing, image correction, stereo matching and other operations on the collected seedling image sequence to obtain the spatial point cloud information of the seedlings; finally, the three-dimensional model of Masson Pine seedlings is obtained after reconstructing the surface of the spatial point cloud by using the triangulation algorithm; and then the relevant SVR calculation is used to calibratte the accurate morphological indexes of Masson Pine seedlings quickly and non destructively.