Original paperEvaluation of ultrasonic sensor for variable-rate spray applications
Research highlights
▶ Typical outdoor conditions had insignificant effects in canopy detection of an ultrasonic sensor. ▶ Depositing water on the transducer of the sensor had a significant effect in canopy detection. ▶ Limiting ultrasonic signal pathway reduced interference between synchronized sensors.
Introduction
Tree liners are young trees grown in nurseries prior to being transplanted to fields or containers where they continue growing into larger, market-ready shade trees. The liners are usually one to three years old, and grow very fast. For example, red maple, Acer rubrum (Autumn Blaze) liners can grow more than 1 m during a single growing season (Mathers et al., 2005). Demand for tree liners by consumers is very strong. The state of Ohio alone annually purchases approximately $14 million worth of liners from other states (Pollock, 2005). Pesticide application plays an important role in maintaining the quality of tree liners by protecting them from potential biological damage. Applicators are supposed to adjust the spray volume as the canopy size changes during the growing season. However, variability in growth rates among liners produces a range of sizes in the field. Thus field adjustment to achieve a reasonable match between sprayer application rate and canopy size is impractical for current liner sprayers. In addition, operators are unable to halt spraying during gaps between trees in a row. Thus, a sprayer that automatically adjusts spray volume based on sensed canopy size is anticipated to achieve two goals: to maximize efficiency by applying the optimum amount of spray to target trees and controlling spray between trees.
Measuring canopy size is a challenge due to the complicated growth structures and irregular shapes of trees. Various remote sensing techniques have been investigated to achieve this goal. For example, light interception and aerial stereoscopic imaging techniques have been adapted to estimate tree canopy size (Meron et al., 2000). Satellite images have also been used to estimate canopy volume of trees in forest (Mäkelä and Pekkarinen, 2004, Carreiras et al., 2006, Mõttus et al., 2006, Le Maire et al., 2008).
However, the scale of these remote sensing techniques is relatively large and consequently, the sensing resolution may be insufficient for a real-time variable rate application in a liner production field. In addition, remote sensing techniques typically have a chronological gap between detection and application, resulting in application errors. To reduce this problem, a LIDAR (Light Detection and Ranging) system or a laser scanner has been used to measure canopy volume. Promising results were reported for using this system in which measured canopy volume was close to manually measured volume (Wei and Salyani, 2005, Lee and Ehsani, 2008, Rosell Polo et al., 2009). Unfortunately, the narrow row spacing in a liner field may restrict LIDAR from being used on variable rate liner sprayers. It is also a relatively expensive sensor ($2000–6000). Furthermore, a typical tree liner sprayer treats multiple rows at a time. Each liner row would require an individual LIDAR system to measure its tree canopy variation. Thus controlling a variable-rate application sprayer would require several LIDAR systems. This would increase the application cost to an impractical level.
Ultrasonic sensors that are affordable, relatively robust during outdoor conditions, and capable of estimating the canopy volume of trees satisfactorily have been used by several researchers (Giles et al., 1988, Tumbo et al., 2002, Zaman and Salyani, 2004). These studies were focused on sensing canopy volume of fully grown orchard trees with relatively large spacing between rows that provided sufficiently clear line-of-sight for the sensors. Their field conditions were different from a typical tree liner field: typical liner spacing is from 1.22 to 1.52 m. Furthermore, during a growing season, row spacing became even narrower due to canopy development. Therefore, canopy detection methodology suitable for an orchard may be inappropriate for liner applications.
While ultrasonic sensors have been used in earlier studies for detecting canopy size, sensor performance has not been well examined under field conditions. In addition, liner field application presents unique challenges to a canopy sensing system, i.e., relatively dense tree liner arrangement, rapid canopy size and color changes, and limited working space between rows for the sensing system (Fig. 1). Ultrasonic sensors may overcome these challenges due to their small size, robust sensing mechanism against color variation in targets, and uni-directional sensing line-of-sight.
However, although the performance of ultrasonic sensors have been presented in the literature, questions regarding the sensor's performance under harsh field, spray application, and multiple sensor operating conditions still remain. For example, Zaman and Salyani (2004) reported canopy density and ground speed influence on tree canopy detection when using ultrasonic sensors. Although they reported on ground speed and foliage density effects on sensor measurements, they examined the sensor performance under limited field conditions and for relatively slow travel speeds (1.6–4.7 km/h) compared to typical liner applications. In addition, Giles et al. (1988) reported that traveling at the speed of 2–6 km/h had no significant effects on the capability of their ultrasonic sensors to detect tree canopy volume. The performance of multiple sensor operation should be evaluated while they are operating simultaneously because measuring tree canopy requires multiple sensors (Giles et al., 1988, Tumbo et al., 2002, Zaman and Salyani, 2004, Solanelles et al., 2006, Gil et al., 2007). Although multiple ultrasonic sensors were used in one system to detect tree canopy, sequentially triggered sensors were used in their study to prevent interference between adjacent sensors (Tumbo et al., 2002, Zaman and Salyani, 2004). However, sequentially triggering sensors are not a feasible option for liner application due to rapid changes of liner canopy sizes. Consequently, the reported results were not sufficient to conclude that an ultrasonic sensor was feasible for tree liner field spray application.
Therefore, the work presented here was to test an ultrasonic sensor for field sprayers, particularly in tree liner application, with a possible contribution toward developing a testing protocol for outdoor-use ultrasonic sensors. A wide range of parameters were evaluated in this study to determine sensor performance under a wide range of field conditions. Parameters studied included: cold weather exposure, cross-wind, dust environment, air temperature, spray cloud and multiple sensor operation. Therefore, the overall objective of this research was to verify the feasibility of using an ultrasonic sensor for tree liner field sprayers. The specific objectives were:
- (1)
to test the durability and measurement stability of an ultrasonic sensor under laboratory simulated, potential field spray application conditions and
- (2)
to determine the optimum sensor implementation strategy for a variable-rate tree liner sprayer.
Section snippets
Ultrasound sensor
An outdoor-use, water proof ultrasound sensor (LV-MaxSonar-WR1, Maxbotix Inc, Brainerd, MN, USA) was used in this research. The sensor was rated as IP (ingress protection) 67 which refers to dust tight (6) and 1-m water immersion protection (7) (CENELEC, 2000). The sensing resolution was 3.82 mV/cm with an approximate beam angle of 10°. The sensor body was constructed with a pipe connector and cable grip to protect the sensor under the outdoor conditions (Fig. 2).
Although other ultrasonic
Cold weather exposure
There was no change in function or accuracy of the ultrasonic sensor after it was exposed to outdoor cold weather conditions for 40 days. The RMS errors of the measurements ranged from 2.15 to 4.06 cm with the mean of 3.31 cm, and 2.71 to 4.94 cm with the mean of 3.55 cm for before and after the exposure, respectively (Table 2). The increase of the mean RMS error was 7.3% after the exposure; however, ANOVA results indicated that the increase was not significant (P > 0.05). RMS errors from the test
Summary and conclusions
The durability and measurement stability of an ultrasonic sensor were investigated under simulated field conditions. In addition, potential issues in detecting a target with multiple synchronized sensors were investigated by integrating them into a prototype sprayer. Although the sensor showed an inherent issue in distance measurement accuracy when the spray liquid deposited on the transducer, the error could be minimized by optimizing the sensor/nozzle relative mounting locations on a sprayer.
Acknowledgments
The authors express their appreciation to Dr. Robert D. Fox for his thoughtful review, and Adam Clark, Barry Nudd and Keith Williams for their technical assistance. This research is supported by USDA NIAR SCRI.
References (20)
- et al.
Estimation of tree canopy cover in evergreen oak woodlands using remote sensing
Forest Ecology and Management
(2006) - et al.
Variable rate application of plant protection products in vineyard using ultrasonic sensors
Crop Protection
(2007) - et al.
Calibration and validation of hyperspectral indices for the estimation of broadleaved forest leaf chlorophyll content, leaf mass per area, leaf area index and leaf canopy biomass
Remote Sensing of Environment
(2008) - et al.
Estimation of forest stand volume by Landsat TM imagery and stand-level field inventory data
Forest Ecology and Management
(2004) - et al.
Estimation of crown volume for a geometric radiation model from detailed measurements of tree structure
Ecological Modelling
(2006) - et al.
A tractor-mounted scanning LIDAR for the non-destructive measurement of vegetative volume and surface area of tree-row plantations: a comparison with conventional destructive measurements
Biosystems Engineering
(2009) - et al.
An electronic control system for pesticide application proportional to the canopy width of tree crops
Biosystems Engineering
(2006) 2009 Spray Bulletin for Commercial Tree Fruit Growers
(2009)Environmental effects on the speed of sound
Journal of Audio Engineering Society
(1988)European Standard EN60529: Degrees of Protection Provided by Enclosures (IP Code)
(2000)
Cited by (77)
Proximal sensing for geometric characterization of vines: A review of the latest advances
2023, Computers and Electronics in AgricultureField evaluation of a deep learning-based smart variable-rate sprayer for targeted application of agrochemicals
2023, Smart Agricultural TechnologyDifferent sensor based intelligent spraying systems in Agriculture
2020, Sensors and Actuators, A: PhysicalCitation Excerpt :The first results showed that, by adjusting the application rate to the features of the target, the CIS allowed to maximize the quantity of spray applied to the leaves in always higher volumes [64]. In [65] to determine the durability and precision, an ultrasonic sensor simulated the environmental and operating conditions. Test conditions include prolonged cold, outside temperature, crosswinds, temperature changes, and movement speed and effects of the spray.
Foliar deposition and coverage on young apple trees with PWM-controlled spray systems
2020, Computers and Electronics in AgricultureCitation Excerpt :However, these sensors have a low measurement resolution and their precision depends on the forward speed and environmental parameters. Consequently, they cannot take continuous three-dimensional measurements due to large divergence angles and can only be used for point detections (Jeon et al., 2011b). Other sensors include the light detection and ranging (LIDAR) sensor technology.
ECOMSNet – An edge computing-based sensory network for real-time water level prediction and correction
2020, Environmental Modelling and SoftwareCitation Excerpt :For example, the bottom slope along the flow direction, one of the parameters in the DSM model, is uniform in the lab channel but not in the field experiment. The ambient conditions such as cross-wind speeds and temperature have significant impacts on the performance of ultrasonic sensors (Jeon et al., 2011). Finally, the DSM model used in this ECOMSNet was efficient for GVF flow conditions.