|Sudduth, Kenneth - Ken|
Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/21/2006
Publication Date: 2/9/2008
Citation: Sudduth, K.A., Chung, S.O., Andrade-Sanchez, P., Upadhyaya, S.K. 2008. Field comparison of two prototype soil strength profile sensors. Computers and Electronics in Agriculture. 61:20-31. Interpretive Summary: Precision agriculture aims to minimize costs and environmental damage caused by agricultural activities and to maximize crop yield and profitability, all based on information collected at within-field locations. Soil strength, an indication of compaction, is a factor that can vary considerably within fields and can also greatly affect crop yields. Because of this, farmers need a quick and inexpensive way to measure compaction, such as could be provided by an on-the-go sensor. This research compares two new on-the-go sensors that can take measurements continuously at multiple depths while traveling across a field. We conducted tests in two fields and found that the on-the-go sensors performed well and gave comparable results. We also compared our sensors to the cone penetrometer, a device commonly used to measure compaction. The sensors both showed similar trends in compaction to those shown by the penetrometer. However, sensor measurements were easier to obtain and resulted in maps with more detail than the penetrometer measurements. Both sensors show promise as potential tools for improved compaction measurement. The results of this study will be useful to scientists and engineers seeking to improve on-the-go sensor design and to develop efficient test methods for evaluating sensor performance.
Technical Abstract: Soil compaction that is induced by tillage and traction is an ongoing concern in crop production, and also has environmental consequences. Although cone penetrometers provide standardized compaction measurements, the pointwise data collected makes it difficult to obtain enough data to represent within-field variability. Moreover, penetrometer data exhibit considerable variability even at a single location, requiring several measurements to obtain representative readings. For more efficient data collection, on-the-go compaction sensors that obtain data at multiple depths are being developed by several research groups. The objective of this research was to evaluate and compare the field performance of two of these compaction sensors. Tests were conducted at two central Missouri field sites, with soil types ranging from sandy loam to clay. The Soil Strength Profile Sensor (SSPS) measured compaction to a 50-cm depth on 10-cm intervals, while the Soil Compaction Profile Sensor (SCPS) also used five sensing elements and obtained data to 40.6 cm on a 7.6-cm interval. Cone penetrometer measurements of compaction were obtained at intervals along each transect for comparison. Data were compared between the two on-the-go sensors and were also related to penetrometer and soil property data. The repeatability of SCPS data was somewhat better than that of SSPS data, perhaps due to larger sensing elements and/or higher resolution at shallow depths. Data from the two sensors were linearly related, with similar relationships found for each individual site and for both sites combined. The agreement between SCPS and SSPS data (r sq = 0.56 over all sites and depths) was much better than between sensor and penetrometer data (r sq = 0.19 to 0.20). Maps of SCPS and SSPS data for a 13.5-ha field site showed very similar patterns. Maps of penetrometer data were also similar to those of on-the-go sensor data but showed less spatial detail. Variation in soil strength appeared to be primarily related to variations in soil physical properties (e.g., texture, water content). Due to the similarity between SCPS and SSPS data, we conclude that measurements obtained with the two on-the-go soil sensors were affected similarly by soil strength variations within the study sites. Side-by-side comparison of the on-the-go sensors provided a convenient approach to validate sensor performance. The study also provided information to improve on-the-go sensor design and to relate sensor data to other measures of soil compaction.