The thermal behaviour of an orchard is intrinsically related to the plant physiological status and it is commonly observed using thermal imagery, in most cases, provided by a drone or by a satellite. Such remote sensing methods are currently popular since they allow to analyse large amounts of land data with few sensor readings. However, they are restricted by the spatial resolution of the images, which always correspond to top views of the canopies. The latter does not allow for a side recording or analysis of the orchard. In this work, we design and evaluate a portable ground-based system for a manual thermal and geometrical characterisation of an orchard, merging thermal images with LiDAR-based range readings in order to obtain a 3D thermal reconstruction of the crop to overcome the previously mentioned issues. The proposed system can work in Global Navigation Satellite System (GNSS) denied environments and delivers multiple views of the orchard, offering the user a three-dimensional view of the thermal behaviour of the grove. Further, the implemented algorithm classifies points from the LiDAR measurements which correspond to the canopy using a supervised classifier. Later, a matching procedure is performed between such points and the thermal information provided by the thermal camera. In order to reconstruct the entire orchard or only a section of the grove, several frames are registered using the Iterative Closest Point algorithm. The system was tested in two conditions: in laboratory and in field within a plantation of Hass avocado, which is one of the main fruit trees growing in Chile, and its performance is compared with an LI-6400 Infra-red Gas Analyser (IRGA) portable photosynthesis system (LI-COR, Lincoln, NE). (C) 2016 lAgrE. Published by Elsevier Ltd. All rights reserved.