During certain periods of the growing season, potato growers must regularly monitor the nitrogen status of their crops in order to apply fertilizer in the most efficient way.
A common practice is to collect leaves from plants in each field and then send them to the lab for nitrate analysis. Within a few days, growers receive results indicating whether more nitrogen fertilizer is needed or if the performance is normal. The system works, but this process can be accelerated, says I. Wang, docent University of Wisconsin-Madison, Department of Horticulture.
“Collecting leaves takes a lot of time and effort,” says Wang.
“And sometimes the results can be misleading, because the amount of nitrate in the leaves can be influenced by many factors, such as weather conditions or the timing of sampling. In addition, the results do not take into account spatial differences [nitrogen requirements] within the field.”
Project funded USDA National Institute of Food and Agriculture, involves the collection and processing of data from a hyperspectral camera. It is installed on a UAV (unmanned aerial vehicle) or a low-flying aircraft that flies over the studied potato areas.
Wang's team is developing computer models to link images to in-season plant nitrogen status, yield, quality, and end-season economic returns.
“My staff and I hope to develop an online program that will convert hyperspectral images into information about when and how much to fertilize so that growers can maximize profits with minimal environmental impact,” says Wang.
“Factors that cause changes in the state of the canopy, such as nutrient status, the presence and absence of moisture or disease, are associated with spectral reflectance and therefore can be visualized in hyperspectral images,” says Trevor Crosby, a graduate student in Wang's lab.
In a single flight over a 70 by 150 meter research field, dozens of images can be collected, each containing hundreds of spectral bands. To speed up image processing, Wang hired two key people. Phil Townsend, Professor of Forest and Wildlife Ecology, is a leader in remote sensing technology. Paul Mitchell, Professor and Specialist in the Department of Agricultural and Applied Economics, conducts an economic analysis from which a computer model makes recommendations for nitrogen application.
Crosby, taking the lead in ground measurements, collected data from field survey sites at various stages of potato growth. This includes the leaf area index, the total nitrogen concentration in leaves and stems, the number of tubers and the weight of individual tubers, and environmental factors such as soil moisture and temperature, solar radiation, and wind speed. At harvest, it measures the overall yield of tubers and their size.
Crosby then developed improved models linking hyperspectral images to ground based measurements. The goal is to predict the nitrogen status of crops in real time and predict the yield of tubers at the end of the season. At this point, the field work and image processing is complete, and Crosby is focusing on model development.
Wang widely shares his research with the state's potato and vegetable growers. He has a good relationship with farmers across the state and many are looking forward to the results of his research.