Sensorray

• GIF

  Greenhouse Sensor Data

Inspiration

Greenhouses offer farms or families in Maine the ability to grow plants year-round, as long as they are kept within proper conditions for the chosen crops. This can prove difficult, however, as the heat and soil moisture are highly dependent on the amount of daylight for a given day. With such large variability in conditions, keeping a close eye on the plants’ environment is crucial, as is intervening in time to prevent plants drying out or the soil’s pH from varying too far from the plants’ ideal conditions. Sensorray offers a solution in an inexpensive, modular array of sensors that provide intuitive, real-time spatial visualizations of the conditions within your greenhouse at the push of a button.

What it does

Heat map is a very informative visualization of various continuous phenomenon in the greenhouse. Since plants grown in greenhouses are typically sensitive to variations in temperature, soil moistures and other parameters, it is important to visualize and alert this so the farmer can monitor their farm in real time. A typical use case would be to water the plants either manually or automatically once it is seen that the soil moisture content is low. It is also possible to identify leaks and equipment malfunction to take preventative action before catastrophic failure.

How I built it

We identified key components and supporting technologies that already exist in the industry. We chose google cloud Big Table as our primary Data storage solution. Sensor data is inherently temporal and a medium scale sensor network generate large streams of data which need a datastore that can scale vertically while maintaining low latency in bulk read and write operations. To serve our end product i.e the heat map visualization we needed to perform many range scan operations and the way BigTable stores this data is ideal. Sensors are interfaced to Arduino and it uses serial communication to communicate with an end node. This end node can be a raspberry pi, Qualcomm dragonboard 410C or similar computing architectures. For ease of development we chose for serial communication with Python running on PCs. The console application we built on python collect serial data from Arduino and insert it in Cloud Big table hosted with google. A separate application will query the data from this remote database and visualize it as a heat map that updates in real time.

Challenges I ran into

Working with sensors with no serial number and datasheet. Some sensors have very little presence online and DIY communities often have no mention of them. Querying BigTable from Python and designing the schema. We are used to Relational Databse and the learning curve appeared steep. Perhaps it can be attributed to very less sleep.

Accomplishments that I'm proud of

The inclusion of the spatial “heatmap” of the various sensor data is the accomplishment I am most proud of. Being able to take the data from the probes in the soil, transmit them into the google cloud, and then to have them appear on a webpage is an accomplishment I never thought I would actually be able to do. Granted some of these were more theory than achievement, but the idea that if we had more time we would actually be able to do these things is (in my opinion) mind-blowing.

What I learned

I realized that agricultural issues in Maine are far-reaching and complex, and that the many interventions and inventions that people are implementing are a showcase of the ingenuity that comes from necessity.

Built With

• arduino
• dragonboard-410c
• google-cloud
• python
• sensors