Multi-tree woody structure reconstruction from mobile terrestrial laser scanner point clouds based on a dual neighbourhood connectivity graph algorithm

A process is presented for the vector reconstruction of fruit plantations based on the model developed by Verroust and Lazarus. To solve occlusion problems, the use of a dual graph of local and extended connectivity is proposed. The process allows vegetation variables such as the length and volume o...

Descripción completa

Detalles Bibliográficos
Autores: Méndez, Valeriano, Rosell Polo, Joan Ramon, Pascual Roca, Miquel, Escolà i Agustí, Alexandre
Tipo de recurso: artículo
Estado:Versión aceptada para publicación
Fecha de publicación:2016
País:España
Institución:Universitat de Lleida (UdL)
Repositorio:Repositori Obert UdL
OAI Identifier:oai:repositori.udl.cat:10459.1/58099
Acceso en línea:https://doi.org/10.1016/j.biosystemseng.2016.04.013
http://hdl.handle.net/10459.1/58099
Access Level:acceso abierto
Palabra clave:Multi-tree reconstruction
LiDAR
Mobile terrestrial laser scanner
Point cloud
Tree training
Ligneous structure
Descripción
Sumario:A process is presented for the vector reconstruction of fruit plantations based on the model developed by Verroust and Lazarus. To solve occlusion problems, the use of a dual graph of local and extended connectivity is proposed. The process allows vegetation variables such as the length and volume of the ligneous structure to be measured, enabling studies such as intensity of pruning operations. The process has been tested against simulated models and real trees with different training systems: open-vase system (peach trees) and central leader hedgerow system (pear trees). The cost of the algorithm will be given by the cost of the implementation of Dijkstra's algorithm, which in its standard version is of potential (O(n2)). Algorithm accuracy was checked against point clouds of virtual trees. The reconstruction was also applied before and after a pruning operation of real trees to enable a study of the evolution of the vegetation indices. Results showed the algorithm to be suitable for multi-tree reconstruction of both central leader and open-vase training systems.