Agriculture and Computer Science

Computer Science Honours student Yonghong Chen developed an app that can be used to predict the amount of crop yield that an agricultural field will produce using only a picture. This startlingly accurate system is inexpensive and meets a huge demand within the farm industry. Since farmers often have to wait to measure their crops after they have been harvested, this technology will help them make important decisions sooner. A key component of the app is that it allows a user to take a picture of a section of produce, and gives the user the number of crops in the section based on that picture. The development is known as the “Estimage” system. Interestingly, it was developed by first placing coins on a table and asking the app to count the number of coins. This system was also used to count the number of logs stacked in a pile. Eventually it was able to count the number of blueberries on a bush, as well as other agricultural applications. The system is very effective and saves a lot of time and money in the amount of effort it takes to count objects. This clever app combines counting and agriculture in a new and ground breaking way. The surprisingly simple, yet previously underdeveloped idea, has many other features as well. The Estimage system consists of an Android client app for interacting with users, a PHP server app for handling requests, and an Octave program for image normalization. It also consists of an open-source ML software package ilastik that is used to apply a predictive model to an – -image. The Estimage system is very good at detecting shape, color, and size, and is also good at distinguishing between backgrounds and objects, provided that the background is similar to that which was used to train the model.