Autonomous robots shipped with on-board sensors have become an effective remote sensing tool in agriculture. Aerial robots have been mostly used for image surveying, and ground robots have been employed with different sensors and actuators. The sensing range from one robot sometimes is not enough. Because of it angle of view or sensor limitations. Therefore, a fleet of robots that provide heterogeneous sensing information could aid to overcome these limitations.
Apple producers have to monitor their orchards daily within a determinate time window over the season in order to take actions that will reflect the quality of their product and economical income. An important task is flower thinning where the producer decides the amount of chemical that he has to apply in each individual tree. This task is hard to achieve in a reasonable time and is often not accurate enough which is critical for the producers because it can produce severe economic losses.
This work aim to use computer visions and machine learning techniques for automated estimation of flower blossom in apple orchards. This thesis will be carry out in the follow steps: a) Review previous works in machine vision and learning techniques applied to blossom estimation; b) Design a novel approach to solve this problematic; c) Experiments with the already available dataset.
Dias, P. A., Tabb, A., & Medeiros, H. (2018). Apple flower detection using deep convolutional networks. Computers in Industry, 99, 17–28.
- UAV enthusiast
- Willing for learning new software tools
- Excited to work in robotics
Theme(s): Sensing & measuring