ENVI Helps Facilitate Peanut Crop Health in Australia
Growing peanuts is a delicate and precise process. Peanut plants do not flower and dry off as they mature, like wheat or corn, but continue to flower indeterminately and grow their pods underground. This makes it extremely difficult to surmise when they are ready to be harvested, which has resulted in a $6.3 million annual loss to the Australian economy. In addition, many crops can be lost to contamination by aflatoxin. To combat these issues, farmers and researchers needed a way to determine optimal harvest times, and to detect aflatoxin in dry-land peanut crops.
Ph.D. Scholar Andrew Robson analyzes QuickBird Imagery using ENVI to determine peanut maturity and the presence of aflatoxin, a mold that contaminates seedlings of peanut plants. Using a combination of techniques, Robson is able to identify areas of stress in peanut crops, preventing further aflatoxin contamination by allowing farmers to focus their treatment efforts. Robson’s techniques also provide the capability to produce harvest maps, which are used by farmers to maximize crop yields.
In his satellite image analysis processing, Robson uses ENVI to perform point registration, using both image-to-image and image-to-map capabilities to pinpoint target regions of interest (ROIs) within crops and index those ROIs using NDVI and SAVI. After subsetting the ROIs, Robson performs classified and unclassified analysis as well as density slicing to identify areas within each ROI for peanut maturity and aflatoxin existence. The results of the analysis of the satellite imagery proves high correlations between classified NDVI image color zones with peanut maturity, yield and stressed areas prone to aflatoxin contamination.
ENVI allows Robson to perform change detection between sequential images and highlight areas of significant change. Analysis of seasonal images and change detection results allow farmers to assess which areas reflect disease-related canopy loss and focus treatments on those areas to prevent further spread. Final images in effect become a harvest map, enabling the grower to form a harvest management regime that maximizes yield and quality, while minimizing prolonged disease exposure.
Robson employs field spectroscopy and Near Infrared Light (NIR) spectroscopy to identify the level of kernel maturity in target areas, thus allowing him to determine optimal harvest time. Since spectral signatures differ according to kernel maturity levels, he can classify and map wavelengths in canopies, kernels and shells in three maturity grades - immature, mid-mature and mature. Robson has been able to identify specific wavelengths from the peanut canopies that correlate with peanut maturity using ENVI. With ENVI’s spectral library builder, he converts .txt files to spectral library files to view data, perform data smoothing and identify spectral ROIs.
Looking to the Future
Robson’s application has been well received by peanut growers throughout the South Burnett, Queensland area, and has been tested in five other cropping regions in Australia. The success of using satellite and spectroscopy imagery and analyzing it with ENVI has led to plans to commercialize the solution for future crop years. His resourceful discovery of the correspondence of unique wavelengths to peanut maturity will result in constructing LED sensors to be mounted on irrigation booms or tractors, allowing farmers to produce low-cost maps throughout the growing season. Similarly, the application of spectral signatures and their connection to aflatoxin could lead to developing in-line sensors for quality assurance in Australian peanut mills.
- Using ENVI, farmers and agriculturists can now harvest peanuts at a more optimal time.
- Since harvesting crops too late can cause losses in the millions of dollars, this solution saves farmers a significant amount of money on lost crops.
- ENVI's built-in functionality allowed researchers to do the analysis they needed without the need for third party products. It is an all-inclusive solution.