3D Apple Detection from Large Point Clouds Using Deep Learning
In this work, the 3D deep learning object detection model, PointRCNN, is evaluated for apples detection. Point clouds obtained from two different methods, LiDAR and photogrammetry, and a combination of the data obtained from them are used to see the effect of positional accuracy, cloud density and t...
| Autor: | |
|---|---|
| Formato: | tesis de maestría |
| Fecha de publicación: | 2023 |
| País: | España |
| Recursos: | Universitat Autònoma de Barcelona |
| Repositorio: | Dipòsit Digital de Documents de la UAB |
| Idioma: | inglés |
| OAI Identifier: | oai:ddd.uab.cat:285205 |
| Acesso em linha: | https://ddd.uab.cat/record/285205 |
| Access Level: | acceso abierto |
| Palavra-chave: | 3D Point cloud Photogrammetry LiDAR Object detection Deep learning Computer vision |
| Resumo: | In this work, the 3D deep learning object detection model, PointRCNN, is evaluated for apples detection. Point clouds obtained from two different methods, LiDAR and photogrammetry, and a combination of the data obtained from them are used to see the effect of positional accuracy, cloud density and the effect of additional parameters such as color and reflectance. The experimental results in this thesis show that photogrammetry data with color information yielded scores much higher than the scores obtained from the LiDAR data. The average AP score for photogrammetry and LiDAR data is 67.62 % and 26.31 % respectively |
|---|