Cherry tomato maturity identification method and device based on quantitative nuclear magnetic image

Fecha de publicación: 09/08/2024
Fuente: WIPO "processing tomato"
The invention belongs to the technical field of artificial intelligence, and relates to a cherry tomato maturity identification method and device based on a quantitative nuclear magnetic image. According to the method, relaxation information of six internal structures is obtained by using sample multi-echo image data, inter-voxel noise errors caused by voxel-by-voxel processing are filtered, inter-structure relaxation differences are reserved, data complexity is reduced, and structural data precision is improved. A typical six-structure template of a cherry tomato fruit is obtained through a deep learning structure segmentation model obtained through batch data acquisition, image structured segmentation is performed on scanned fruit internal multi-echo image data on the basis of the typical six-structure template, and multi-echo data of each structure (an average value of six structures of each echo) is obtained. On the basis of a chaos particle swarm inversion algorithm, single-index and three-index inversion is carried out on echo data of each structure, and the overall and three-component moisture state of each structure is presented, so that a very accurate maturity judgment result can be obtained, and the method has important practical significance for guiding technical improvement of links such as storage and fresh-keeping, cold-chain transportation and the like and improving the quality of fruits and vegetables.