As the world population grows, climate change continues and reliable human labour becomes increasingly expensive, we must find more efficient and sustainable ways to grow food and sustain ourselves. The UBC AgroBot team will be building a fully autonomous robot capable of analyzing its environment and performing targeted weeding, fertilizing and soil analysis though the use of advanced robotics, image recognition and machine learning.
We intend to use the Tensorflow object identification module, training and fine-tuning our own recognition model for crops and weeds
We will be engineering the specific mechanics of the AgroBot, designing the chasis and exterminating mechanism, ultimately integrating them with the Robot Operating System
We will utilize the latest technology in machine learning to help us construct an autonomous image recognition system and a navigation system
This project will be competing in the AGgrowBOT challenge at Purdue University, Indiana. This competition attracts many teams from companies and universities across North America. The objectives of the competition are to create an autonomous machine capable of navigating a wheat field to identify weeds and analyze crop health.
Upon identification of the wheat plant, the machine should determine if the plant is healthy or in distress which requires fertilizer to be applied by the machine. Upon identification of a weed, the machine should eradicate the weed chemically and/or mechanically. The competition will be judged by reputable members of the agricultural technology industry, such as The Climate Corporation (Monsanto), and Blue River Technology.
The navigation team is dedicated towards the realization of an automatic navigation and positioning system for the robot
The image recognition team works with algorithms closely, in order to train an efficient and accurate crop recognition model
The chassis team will design the basic frame that the AgroBot runs on, giving AgroBot the ability to steer and maneuver
The exterminating mechanisms team devises the optimal strategy to exterminating weed and fertilizing crops, minimizing chemical use
AgroBot reaches out to a wide range of audience through social media, outreach events and competition. Posts will be made on a continuous basis on project updates and competition status while providing coverage for our partnering sponsors. Participation in the agBot Challenge will also provide us and our sponsors great exposure to the agricultural and tech sector in academia, and industry.
Do you have any specific interests or inquiries? Send us a email directly to talk with us!
Subscribe to get all latest news from us.