| Agriculture Remote Sensing: Part Two
Bridging the barriers
By: Mark Servilla The most common forms of remotely sensed data that are used for agricultural purposes are weather information supplied by satellites, and ground-based radar and digital imagery derived from satellite and aircraft-based platforms. Real-time weather information has been widely accepted for many years by the agricultural community as a key tool for managing crops. Remotely sensed digital imagery, however, has not been as successful. The proposed uses of remotely sensed digital imagery include crop stress detection, managing fertility programs, maintenance and monitoring of irrigation systems, damage assessment for insurance adjustment, and many more yet undiscovered innovations. Results from market surveys have shown that there exists a number of barriers that must be bridged before digital imagery will be accepted as a component in regular crop management practices. These barriers are characterized by issues that involve the timeliness, ease-of-use, value, and cost of remotely sensed images. Cost
The economic value of remotely sensed digital imagery to the farmer has yet to be fully demonstrated. Lacking, is research that shows how different applications of digital imagery can offset the cost of farm management. For the economic benefits to be realized, the cost of remotely sensed digital imagery must be in a range where the farm manager can speculate on the economic returns. While digital imagery can be used to pinpoint more efficient applications of fertilizer, herbicides, and insecticides, as well as optimize field scouting activities, commercial remote sensing will not achieve the envisioned wide-spread adoption in agriculture until it provides a proven asset. Value
Perhaps the most difficult question to answer for farm managers is what value does remotely sensed imagery provide, and what different decisions will be made based on digital imagery? The value to digital imagery can depend on the level of sophistication that is used to collect the information. In general, remote sensing platforms that provide the highest quality of data are going to provide the most value to the farm manager and have the greatest cost. To those managers engaged in site-specific farm management practices today, the high quality data will provide the most useful information. This type of data is categorized as "quantitative" and can be used for monitoring seasonal-variable soil/crop conditions.
Data that does not fall into this category and can be used for applications such as mapping soil and crop anomalies, and for providing a quick assessment of crop damage is classified as qualitative. Qualitative data should not be neglected by the farm manager, it generally is the least expensive to obtain and is the most timely due to its non-complex nature. The value of the information provided by a remote sensing platform can be viewed in terms of the spectral capabilities of the sensor (spectral resolution), the geometric properties of the imagery, the spatial resolution of the imagery, and the temporal resolution of the imagery. High spectral resolution sensors can provide more detailed information with regard to differentiating the levels of solar radiation reflected by the crop canopy. Ease-of-Use
For remotely sensed digital imagery to be of use to farm managers the data must be represented in a format that is informative and easy to understand for responsive decisions be to made and actions to be taken. The advent of color infrared photography presented to the farm manager a new method of viewing crops from above. Although this form of imagery provided greater information content on the state of the crop, the farm manager was required to learn the meaning of color shades in the photograph and relate them to variations in their field to recognize the potential use of the photograph. With any new technical innovation that is introduced into agriculture, information must be provided in an interactive fashion that permits input from both the innovator and the farmer. Digital images can take on various presentation formats that range from traditional "true" color and color infrared images to highly modified color maps where specific colors represent a particular state of crop vigor. One adaptation that is being used throughout the commercial remote sensing community is to use a simple color scheme with a minimal number of colors similar to that of the Doppler radar maps that are commonly displayed on most weather reports. In this format the farm manager focuses on key colors that represent variability in the fields under observation. Subtle differences in fields that can lead to confusion are eliminated from this type of color map. Timeliness
A special need of agriculture production management is the input of timely information. In the case of remotely sensed digital imagery, timely input of information consists of both frequent acquisitions during the growing season and fast post-acquisition information processing and distribution to the farm manager. There are still questions to be answered regarding "how frequent" digital images should be acquired during the growing season. Market research has found that a frequency of between every 10 to 20 days is sufficient to effectively monitor crops. Depending on what period of the growth cycle the crop is in, more frequent monitoring may be required. In most cases, a pre-emergence soil image should be considered as part of the seasonal monitoring.
Although today's remote sensing technology can meet the need for frequent monitoring, inclement weather is the ultimate controlling factor. In the mid-west, weather statistics show that there exists an average of 1 to 2 perfectly clear days every two weeks. On average, a remote sensing platform should be able to acquire an image over the target crop within the optimal frequency. Because weather patterns are never consistent, the ability to acquire digital imagery when a fair-weather opportunity occurs is advantageous. For this reason, a popular belief by many in the industry is that airborne platforms have an edge over the present constellation of commercial satellites.
Post-acquisition information processing and distribution times can vary with the remote sensing platform being used and the level of information that the farm manager requires for decision making. In some instances, such as satellite based imagery, part of the correction is performed on-board of the satellite before the data is transmitted to an Earth ground receiving station. In most circumstances, farm managers require information returned on the order of 48 hours or less from the time digital imagery is acquired. Post-acquisition information processing can include calibration of digital data to a known standard for maintaining absolute values of solar reflectance to geometric rectification for correction of image distortions due to sensor view angle, platform attitude, or target surface features. Greater correction requirements in the post-processing stage increase the time between acquisition and when the information is distributed to the farm manager.
These time-sensitive issues are addressed by Photon Research Associates, Inc. (PRA) of San Diego, California, and Data Transmission Network, Inc. of Omaha, Nebraska, who teamed together to develop ÁVerde!ª, which is capable of providing distribution to growers within 24 hours of acquisition through DTN's satellite information system and the Internet. Benefits from this near real-time delivery of imagery include inspection of irrigation equipment prior to emergence, variable-rate irrigation scheduling, and the detection of misapplied fertilizer.
Another issue that is often ignored is the method in which digital information is distributed to the farm manager. For digital information to be most effective in farm management decision making it should be accessible in a manner that is familiar to the farm manager. Today, the World Wide Web and the Internet are becoming a common information gateway for many farm managers. Distribution of digital imagery in this manner can be effective if and only if the imagery volume is low enough that it does not exceed the limitations of the bandwidth of the local Internet service provider. The internet allows the farm manager flexibility to decide when to retrieve their images and selectivity in what images to acquire. Other options exist for the farm manager, but generally require a subscription to a commercial information provider such as DTN. This type of service is usually more reliable in providing timely information since it does not depend on outside sources for distribution bandwidth. A Working Tool
In an attempt to bridge the barriers to agriculture remote sensing, PRA and DTN have offered their prototype agriculture remote sensing tool, ÁVerde!, for the 1998 growing season to farm managers in Hamilton County, Nebraska. Available at a ground resolution of 2 meters, the aircraft-based digital imagery covers a 520 square mile area in Hamilton County. This qualitative imagery is acquired at 2-week intervals during the growing season for a total of 6 acquisitions. The digital imagery is presented in 3 different formats: true color, color infrared, and a color specific anomaly map that highlights crop variability.
For this select group of farm managers, the 1998 growing season has provided them with first-hand experience on how artificial some barriers can be. With season-long access to low-cost imagery, they have experienced the timeliness, ease-of-use, and value, that remote sensing can offer. The derived benefits of the digital imagery are still being calculated through a number of research farms in Hamilton County. With any success, the participants in this program will become new advocates for this under-utilized agriculture tool. About the author:
Mark Servilla, Ph.D., is a research engineer with Photon Research Associates, Inc. He has 2.5 years applied research experience. His research focuses on applications in agriculture remote sensing. He has also worked on developing data warehousing and distribution software for remote sensing imagery. Back |
