A Critical Look at Crop Imagery in Onion Production
By Kyler Beck, Research Support Scientist, University of Idaho Parma Research and Extension Center
Drones are more readily available to onion growers than ever before. Fields and ditch banks have been transformed into miniature airstrips and landing platforms with drones being used to transport tools such as insect sweep nets and deliver anything from small doses of pesticide to cages full of beneficial insects.
Of primary interest to many in the industry, however, has been the use of drones to carry advanced imaging equipment and collect imagery data. Interest in these tools has facilitated the formation of an entire crop imaging industry within the agricultural sector in recent years so that today the technology is available to nearly everyone.
But despite the widespread availability, it appears that interest from some growers in this technology is diminishing. There could be a number of reasons for this. For example, the cost of hiring someone to collect and process images or to purchase and use the equipment itself may very well be keeping some from trying out these tools altogether. However, some of the disinterest I have witnessed is from those who have tried out the technology but have found that it failed to live up to the expectations of usefulness that had gained their interest in the first place.
Therefore, I want to briefly answer two important questions: what is the role and practical use of crop imagery in onion production at the present time, and what are the factors limiting its current uses?
Role of Crop Imagery
Crop imagery has at least the potential to remotely (from afar) diagnose and map onion disease. Thus, remote sensing research is important to the onion industry because the mapping of plant stress and disease within a field is a prerequisite to precision crop protection. In other words, the projected goal of modern agriculture – applying fertilizers and pesticides only when and where they are needed – leans heavily, if not entirely, on the ability to first specifically and accurately account for the spatial variability of disease.
Exploration of the Technology
At the University of Idaho Parma Research and Extension Center, we are currently in our third year of evaluating remote sensing as a tool to diagnose and map pink root in onions. When measuring individual leaves, root disease does not appear to have a distinguishable impact on the color spectrum of onion, suggesting that disease symptoms are, in some cases, too subtle to be identified. However, when measuring a group of onion plants from overhead, as a drone does, the stunted foliage of root-diseased plants may, if severe enough, result in an altered color spectrum.
To be clear, this change is not unique to root disease and can equally result from abiotic stressors (e.g. drought, lack of fertilizer) or variations in plant growth stage. Therefore, the data suggest that crop stunting overwhelmingly impacts the color spectrum so that remote sensing imagery of an onion crop is primarily an illustration of relative foliar ground cover. For a limited but explicative depiction, see Table 1. This is also the case when imagery data is translated into a Normalized Difference Vegetation Index (NDVI) heat map (Fig. 1) or a heat map of other commonly used vegetative indices.
Present Practical Applications
So, what is the potential of onion imagery as it stands, and how can it be practically used in onion production? Given the overwhelming sensitivity of onion imagery to foliar ground cover, it can be used primarily as a tool to monitor foliar growth. Images acquired at similar points within a growing season in different years may provide an indication of how a crop is progressing in comparison with past years. If one is responsible for many acres of onion ground, imagery may prove to be a useful first step in identifying problematic areas upon which to focus scouting efforts. Furthermore, if particular areas of a field consistently result in smaller onion plants, further efforts can be initiated to determine the cause of the issue through intensive soil sampling. Finally, in cases where herbicide drift damages an onion crop, imagery may be useful in quantifying the area affected for insurance purposes.
Limitations to Advanced Uses
Those interested in using imagery to make crop management decisions must rely on pre-established relationships (developed and validated in scientific studies) between crop reflectance and crop health. This leads us to consider two primary limitations of remote sensing at the present time.
First, subtle differences among onion leaf reflectance, such as those which may indicate disease stress, are overwhelmed by the effect of leaf amount when imaged from above. A group of onion plants that are stressed for nitrogen may be lighter green in color compared to a non-stressed group of plants. However, if the non-stressed group are smaller in size (i.e. slightly later emergence) the larger, yet stressed, group will be identified as healthier using common imagery tools like the NDVI. This challenge will continue to limit the advanced use of imagery unless extremely detailed resolution images can be acquired (< 1 inch pixels) and soil pixels are removed prior to the generation of vegetative index heat maps.
Second, the impact of onion stress and disease upon crop imagery data (crop reflectance) has not been extensively researched. If the overall goal is to collect crop imagery of a field and to be able to map areas of stress and disease, then we must have a good grasp on how all types of stress and disease impact onion imagery. Extensively researching the impact of one disease is surely interesting and does contribute to the overall goal; however, that information is of limited practical use by itself. Until the impact – or lack thereof – of each individual disease which could be present within a growing region is characterized and documented, precision crop protection will remain a goal instead of progressing into a reality.
The Future
Remote sensing as a tool still has limitations. Yet, with these limitations in mind, we are better able to strategize for the future. With regard to future research efforts, conducting a widespread survey of the possible onion diseases in a particular growing region with detailed spectral sensors would be a foundational step in the right direction. In this capacity, a plant diagnostic lab seems to be a naturally suitable place to carry out such efforts.
There is only one thing we can say for sure about the future of crop imagery in onion production: it will be determined by the actions and interests of growers and researchers.
