Home gardeners face yearly issues with frost damage to their plants, which can reduce crop yield or even kill the plants when they are young. Measures such as frost blankets (sheets of fabric or plastic covering crops) can mitigate the effects of overnight frost damage by keeping plants warmer than they would otherwise be, but these measures must be taken before a freeze begins, and blankets must be lifted first thing in the morning. Since government-issued freeze warnings are not always accurate, if an unpredicted freeze occurs and a home gardener has not covered their plants overnight they are liable to suffer crop damage. This leads to both an economic loss for the home gardener as well as an environmental loss to society, since home gardens confer benefits to local insect and bird populations and are a more productive use of space than lawns.
This project aims to develop an inexpensive, open source Internet of Things (IoT) system that enables home gardeners to track the temperature and humidity in their garden and receive automatic alerts when an incoming freeze is likely. This is achieved using an outdoor temperature and humidity sensor connected to a microcontroller that relays the sensor readings to the internet. These components are stored in a waterproof enclosure that can be easily mounted to a post in a home garden. A cloud service is used to take in sensor readings and process them to determine the likelihood of a freeze in the near future. When a freeze is determined to be incoming, the cloud service sends a push notification to a mobile app on the gardener's phone alerting them about the likely freeze so that they have time to cover their crops. Gardeners can also check the current temperature and humidity in their garden at any time by opening the app.
Code has been written, tested, and finalized for the microcontroller and most cloud components of this project, and the endpoint has been shown in testing to reliably sense current ambient conditions. The mobile application is under development for use with iOS devices, along with the push notification and user authentication functions of the cloud service. Once the iOS application is finished, future developments could include using machine learning to improve freeze prediction accuracy, as well as the development of software for systems other than iOS to handle the frost warning notifications, including Android, web, and embedded applications.