Machine learning techniques and satellite data are being used to detect sugarcane diseases early.
Researchers at James Cook University are using a combination of machine learning and satellite data to detect sugarcane diseases before they would normally become apparent.
A team led by Professor Mostafa Rahimi Azghadi developed a software tool and tested its ability to accurately distinguish between healthy and diseased sugarcane.
This is the first use of satellite data to target asymptomatic ratoon stunting disease (RSD).
“RSD can affect sugar yields by up to 60% and is highly contagious. However, it is asymptomatic and cannot be seen with the naked eye until late in the growing season,” Professor Azghadi said.
RSD is typically identified by manually cutting and sampling sugarcane and sending the sample to a laboratory for DNA analysis.
“Each test costs about $10 to $15, so it’s time-consuming and expensive, especially if you want to do it on a large scale,” Professor Azghadi said.
“Depending on the sugarcane species, the accuracy of our method was between 86 and 97 percent, which is comparable to or better than other crop disease detection tools.”
Something like a visit to your family doctor
In this study, various machine learning techniques were utilized to detect the presence of RSD in different types of sugarcane using vegetation indices obtained from freely available Sentinel-2 data.
Scientists arranged to collect veridical samples across 76 sugarcane plots in the Herbert region of Queensland, Australia. This dataset was acquired by trained field agronomists from Herbert Cane Productivity Services.
“The ground truth data was then used to extract 76 sampled blocks from the Sentinel 2 image, ensuring that each pixel within the block geometry was labeled with both disease status and diversity,” the researchers said in their paper.
They found that machine learning algorithms can “effectively classify RSD across several breeds using freely available satellite-based multispectral data.”
“RSD in sugarcane is just the first case where we have succeeded…our approach can be extended to other crops and other crop health problems,” Professor Azghadi said.
“The long-term goal is to develop early warning tools to identify disease risk, track overall crop health, and make it easier for farmers to manage crop health and vigor.”
“It’s a bit like a regular check-up at your GP, but for sugarcane and other crops.”
A promising approach
writing in diary Information processing in agriculturethe researchers said, “Our study highlights the potential of satellite-based remote sensing as a cost-effective and efficient method for large-scale sugarcane disease detection as an alternative to traditional manual laboratory testing methods.”
Highlighting the advantages this method offers when compared to traditional methods of detecting RSD, the researchers go on to state that “freely available satellite-based remote sensing provides a cost-effective and efficient alternative to traditional resource-intensive methods of identifying and managing sugarcane diseases.”
“In particular, free and publicly accessible multispectral satellite data can reduce the financial burden of purchasing expensive spectroscopic planning images and facilitate widespread adoption of this advanced technology.”
“These promising initial results, coupled with the efficiency of classifying 76 blocks within minutes rather than months, demonstrate the potential benefits of implementing large-scale health monitoring systems using satellites and machine learning.”
