While it's important to keep crops irrigated, you certainly don't want to overwater them – particularly in regions where water is scarce. With that in mind, scientists have now developed a method of gauging the moisture levels in soil, using an ordinary camera.
First of all, there are already technologies that allow farmers to see how much water is in the soil, but they're not without their limitations.
There are buried moisture sensors, for instance, which can be adversely affected by salt in the soil. Thermal imaging cameras can also be used, although they're expensive, plus their readings may be thrown off by climactic factors such as fog or cloud cover. Drone-mounted imaging systems have even been used, but the drones themselves are limited by adverse flying conditions and relatively short battery life.
Seeking a simpler, cheaper and more reliable alternative, researchers at the University of South Australia and Iraq's Middle Technical School looked to an ordinary RGB (red, green, blue) digital camera, along with machine learning-based software.
The experimental system was "trained" over a four-week period at an agricultural nursery in Iraq. Doing so involved using a tripod-mounted Nikon D5300 DSLR to take photos of the monitored soil from a variety of distances (1 to 5 m/3.3 to 16.4 ft), at different times of day, in both sunny and cloudy weather conditions.
The moisture levels of the soil had already been measured at the times the photos were taken, using conventional methods. Therefore, the software was able to establish which subtle variations in the coloration of the photographed soil corresponded to which known moisture levels. When the system was subsequently tested, it proved to be highly accurate at estimating soil moisture based on soil coloration in photographs.
"Once the network has been trained it should be possible to achieve controlled irrigation by maintaining the appearance of the soil at the desired state," says the University of South Australia's Prof. Javaan Chahl, who led the study along with Dr. Ali Al-Naji. "Now that we know the monitoring method is accurate, we are planning to design a cost-effective smart irrigation system based on our algorithm using a microcontroller, USB camera and water pump that can work with different types of soils."
A paper on the research was recently published in the journal Heliyon.
Source: University of South Australia
