Hi vertical farmers,
In the previous post we did a walkthrough our web application for supervision and control. In this post we will show our first experiments with the vision system introduced on Post 7 - Vision System.
Just to refresh the concept, our vision system is a set of computer vision algorithms capable to detect plant characteristics.
First, we provide photos with characteristics that we want to detect, for example dry leaves, small sized plants or sick plants.
Then, we run the system normally: we have a camera on the vertical farm taking pictures that are analyzed by the algorithms that search for the desired characteristics.
Follows our first experiments with the feature detector algorithm SURF (speeded up robust features). The left images are inserted in the algorithm, resulting the images on the right:
This is the first stage of the feature detection process. It’s followed by the use of probabilistic methods to extract features of interest.
It can be seen that the algorithm is heavily detecting corners, edges and plants. We don’t have much calibration photos to adjust the the algorithm yet.
Our work proceeds and we hope to have better detections to show soon.
As always if you have any questions or comments please feel free to comment this post.
Thanks for following and keep connected!
Top Comments