B. Engineering (I.C.T. Software) & B. Industrial Design (3D Graphics)
Studios A & B
This was a Professional Studio project I spanned over 2 semesters with 2 separate groups. The project originally focused on computer vision, and machine learning data analysis on data captured by drone of a vineyard to assess the vineyard’s health. The project employed the same principals to updated data for it’s final implementation to a macadamia nut farm.
The Studio subjects are build to give students a real life scenario of working with a client to a brief. Our client for Canopify was Southern Cross Drones. I liaised with Dr Ben Rodanski, the Chief Technology and Science Officer to his brief and data Southern Cross Drones provided.
Click here to see achievements
Negotiated the client brief as per clients wishes communicating with clients, team members effectively. Effective use of Git and Trello for version control and team management providing value in the workplace as per Engineers Australia areas of competency.
By liaising with the client on legal obligations and potential risks posed by Canopify I continued my development in line with Engineers Australia areas of competency.
Build a Machine Learning model which accurately detected plants with poor canopy health practicing competently according to Engineers Australia areas of competency.
Successfully learnt how computer vision works with various methods of training a computer vision machine learning model. These skills added to my technical ability as per Engineers Australia.
Key Topics Researched
Computer Vision
Researching different brute force and morphological processes that can and are applied in various libraries and algorithms for computer vision.
This included manipulation of pixel matrices for better definition within the image en mass before decorrelation and binarizing to identify specific shapes and patterns for processing.
Machine Learning
Finding and experimenting with different methods of training machine learning models. Spending considerable time defining images for the models to be trained and successfully detect unhealthy vegetation canopy.
Team & Project Management
I really drilled down in Git and Trello. I wanted to see what Trello can do and whether the team could work cohesively between the two.
This involved close contact with clients and the team conversing between the two on project concerns. Merging work suitable for the project to implement in production to the master branch for client review and presentation. This included continual review and drilling down on Trello cards keeping the team in line and on time for their due dates.
