Monitoring system of an apple orchard Design, development and implementation of a monitoring system for an apple orchard, through the creation of an integral digital twin

The following thesis project describes the design, development and implementation of a monitoring system for an apple orchard, through the creation of an integral digital twin (DT) of said orchard that supplies useful information about the apples, such as health and position. The work done for this thesis corresponds to the continuation of a project carried out by the Department of Environment, Land and Infrastructure Engineering (DIATI) and the Department of Mechanical and Aerospace Engineering (DIMEAS) of the Politecnico di Torino. Data acquisition was performed using multispectral cameras and a stereo camera mounted on an agricultural UGV developed by DIMEAS, called AgriQ, that was driven around the orchard by remote control. The multispectral cameras were used to measure the light not absorbed by the apples during photosynthesis indicating the health of every single fruit. On the other hand, the stereo camera recorded an SVO file from which a point cloud was extracted and, thanks to the to the integrated Inertial Measurement Unit (IMU) and the possibility of using artificial intelligence (AI) and neural networks, apples were detected and located. In addition, a LiDAR scanner was used to obtain a georeferenced point cloud of the orchard. The project required the design of the support for the sensors to be mounted on the UGV, taking into consideration the orientation and placement of the cameras needed to obtain the desired framing of the apple trees and the corresponding overlap to ensure a correct redundancy of the pictures. After data acquisition, point clouds were extracted from the outputs of the various cameras and sensors and were manipulated in order to obtain a single point cloud. It is interesting to mention the work done to recognise the apples, for an AI model had to be trained to detect apples and then integrated on an application to recognise and locate the apples. The aforementioned materials and procedures are better explained throughout the thesis, together with the used softwares and methods. Furthermore, conclusions, limitations and future applications are also discussed at the end of the thesis.