Technical Summary: 

The proposed project is designed to develop an optical yield monitor system for precision agriculture in pneumatically conveyed crops such as peanut, sugar cane, and forage etc.  The concept for the sensor portion of the monitor is based on the design of the cotton yield monitor sensor developed previously at Mississippi State University.  The actual sensor will require significant modification for each crop on which it is to be used, because optical properties, flow-system design, and contamination considerations are different with each crop.  The concept of the yield monitor system includes an advanced optical mass-flow sensor and a data acquisition system.  The mass-flow sensor detects crop mass by sensing the optical properties of the crop in situ in real-time.  The proposed mass-flow sensor is unique, as the sensor is a single unit.  With a single unit, no alignment is required in sensor installation.  This makes it very easy to install and maintain.  The proposed mass-flow sensor has an anti-stray-light feature.  Ambient light change does not affect its performance, so the sensor can be installed at a position where the sensor’s window can be remained clean.  Thus, the error typically induced by sensor window contamination is alleviated.  This feature will make it possible for the yield monitor to be used for crops that contain large amounts of trash, dirt, and dust in harvesting process.  Furthermore, the proposed mass-flow sensor has a built-in temperature control system.  During operation, the sensor’s inside temperature is controlled to maintain it at a constant point.  This characteristic makes the yield monitor system accurate and stable when it is used under varying temperature environments.  The output signal of the sensor and spatial information from a GPS receiver will be collected and processed by a data acquisition system with algorithms designed for proper data handling and storage.  The proposed yield monitor system is expected to be capable of providing spatial yield information for precision agriculture in pneumatically conveyed crops. 

The goals of this project are (1) to develop a yield monitor system, that includes optical mass-flow sensors and data acquisition and processing units, for pneumatically conveyed crops such as peanut, sugar cane, and forage; (2) to develop a method for installation and operation of yield monitors for pneumatically conveyed crops; (3) to build 3 to 5 prototypes of the yield monitor systems for field testing and evaluation; and (4) to license the intellectual property which the proposed project may generate to private companies for commercialization.

Because crops like peanut and sugar cane are economically important crops in many countries including the U.S., and precision agriculture is being adopted more widely, application of the proposed yield monitors will definitely benefit the users and the economy. 

Objectives: 

The overall goal of this study is to develop an optical yield monitor system for precision agriculture in pneumatically conveyed crops such as peanut, sugar cane, and forage etc.  

Objectives for the first year of this project are:

1. Based on the optical properties of different crops, to design and fabricate optical sensors to detect mass-flow of the crops.
    
2. To design and fabricate a data acquisition system for collecting and processing GPS data and the mass-flow data from the sensor.
    
3. To conduct preliminary laboratory and field tests of the yield monitor system.

Objectives for the second year of this project are

1. To make adjustments and improvements to the yield monitor based on the preliminary test results obtained from previous year.
    
2. After refining the design, build 3-5 sets of the yield monitors for intensive field-testing.
    
3. To intensively field-test the prototypes of the yield monitor.

Procedures: 

The proposed yield monitor includes an optical mass-flow sensor and a data acquisition system.  The mass-flow sensor comprises a single housing unit (Figure 1).  The sensor includes light sources such as light-emitting diodes (LEDs) mounted along the surface of the housing unit.  The light sources each project a light beam into the flow passage in a direction normal to the longitudinal axis of duct. The sensor also includes one or more photodiode detectors mounted along the surface of the housing unit as well.  Thus, light sources and detectors are contained in a single housing unit, mounted in a single location, and facing in the same direction (i.e., in one of side walls at one side of either inlet end or outlet end or outside of duct).  Generally speaking, the mass of the entrained materials is determined by measuring the interaction of light with the materials.  The sensor’s spectral characteristics must be adjusted according to the spectral properties of the measured material to maximize the sensitivity of the sensor.  In optical measurement circuits, such as used in the proposed mass-flow sensor, only the light received back from the entrained materials is measured.  A transparent window is located on the surface of the housing unit between the flow of materials and the light sources and detectors.  The sensor window protects the light sources and detectors from the entrained materials in the duct as the materials pass by the mass-flow sensor.

Figure 1.  Operating principle of proposed mass-flow sensor 

A light modulation technique will be applied in the mass-flow sensor design to give the sensor anti-stray-light capabilities.  Light modulation is accomplished as follows:  LEDs used as light sources for the sensor are modulated by using a pulse generator circuit; the pulse generator creates high frequency pulses; the output of the pulse generator drives transistors to turn on or off the LEDs while the pulses are high or low, respectively. 

The light measurement circuits include a photodiode detector, a current to voltage amplifier, an electronic filter, and signal amplifiers.  The photodiode transforms the light into an electric current.  The current, proportional to the light intensity, is converted into an electrical voltage with a current-to-voltage amplifier circuit.  Output from the current-to-voltage circuit is followed by a high-pass filter, which allows the signal associated with modulated light returning from the material passing through the duct to be detected, while the signal component generated by stray-light is filtered out.  Thus, the signal after the high-pass filter has no significant component that is attributed to natural illumination (sunlight).  A demodulation circuit follows the high-pass filter.  After the signal is demodulated, operational amplifiers then amplify and output an analog signal that is proportional to the intensity of modulated light returning from the material passing through the duct. 

The proposed mass-flow sensor has internal temperature control that greatly improves measurement accuracy when operating under varying ambient temperature.  The temperature control circuit will consists of a temperature sensor, a reference voltage, a voltage comparator, a solid-state relay, and a thermo-electric module.  A temperature control point is determined by the reference voltage circuit. The output of temperature sensor is proportional to the temperature inside the sensor body.  The comparator compares the reference voltage with the output from the temperature sensor.  If the reference voltage is higher than the temperature sensor output, the output of the comparator drives a transistor to turn on a solid-state relay.  In that case, the thermo-electric module is connected to power, and the temperature inside the sensor increases.  When the sensor temperature reaches the temperature control point, and the temperature sensor output becomes higher than the reference voltage, the output of the comparator causes the transistor to turn off the solid-state relay, and the thermo-electric module is thus disconnected from power, so that sensor temperature decreases.  In this way, a relatively constant sensor temperature is maintained. 

The data acquisition system will be developed to record sensor outputs and GPS data in real time based on preset algorithms.  It will include a palmtop computer and an ACD (analog to digital converter).  Yield information will be displayed on a screen and stored in a PCMCIA memory card.  A DGPS will be employed for use with the monitor.  The GSA sentence and RMC sentence from the DGPS receiver will be used to provide PDOP (position dilution of precision), location, and speed data.  Location data are differentially-corrected with the signal from the nearest U.S. Coast Guard beacon station.  The system’s data acquisition box directly reads data from the DGPS receiver.   

In the first year of this project, two prototypes of the proposed yield monitor, including four optical sensors and two data acquisition units, will be developed for preliminary laboratory and field testing. Two crops (probably peanuts and sugar cane) will be the focus of the work in year 1.  In the second year of this project, three to five more prototypes of the yield monitor will be developed for intensive field-testing and evaluation by end users. 

Justification: 

Precision-agriculture technologies provide a way to adjust production inputs based on the needs of individual areas within fields.  These adjustments can be managed to optimize profit and minimize environmental impact.  Optimizing profit requires knowledge of the amount of crop yield at a given point in a field.  Thus, yield monitors are very important for the future of precision agriculture in various crops.  Yield monitors that incorporate GPS data have been successfully implemented in grain crops, and they have been commercially available for several years.  Development of yield monitors for non-grain crops has been slow, partially because of their harsh harvesting environment and inhomogeneous material properties.  So far there is no widely accepted yield monitor for crops like peanut, sugar cane, and forage.  As precision-agriculture technologies have become more and more widely adopted in production agriculture, accurate, reliable, inexpensive, and easy operating yield monitors for those crops are greatly needed by producers. 

So far PIs do not have enough information to draw a conclusion on the economic potential for yield monitors in pneumatically conveyed crops such as peanut, sugar cane, and forage.  However, PIs understand that (1) peanut and sugar cane are economically important crops in many countries including the U.S., and they are high value and high input crops; (2) as precision agriculture becomes more and more widely adopted, farmers will want and need yield monitors because yield is a the most important factor regarding profitability; and (3) there is no such yield monitor commercially available now. 

Literature Review: 

Vellidis et al (2001) developed a Peanut Yield Monitoring System (PYMS).  PYMS used load cells for instantaneous load measurements of harvested peanuts and has proven to be as accurate as within 2 to 3% on a trailer-load.  The instantaneous accuracy of PYMS was calculated to be 700 kg/ha.  The field evaluation showed that basing management decisions on the yield of individual pixels of PYMS yield maps is not realistic.  The strength of PYMS is in differentiating yield trends and evaluating management practices.  

Benjamin et al. (2001) designed and tested a sugar cane yield monitoring system, mounted on a Cameco sugar cane combine.  The system was comprised of a yield sensor, a data acquisition system, and a differential global positioning system (DGPS).  A scale was used as the yield sensor in the system.  The scale was mounted in the floor of the elevator, and it took instantaneous measurements of the cane yield directly.  Field-test results showed that among the 118 tests, the error ranged from 0 to 33%, and 14 of them had an error above 20%.  The average error was 11.05%. 

Wilkerson et al. (1994) developed a sensor to measure real-time cotton flow.  Their work included a light-source array that projected light across a cotton-picker discharge chute.  On the opposite side of the chute was a photo-detector array that measured the amount of light crossing the chute.  Measuring light attenuation caused by passing particles allowed calculations of the amount of cotton passing the sensor cross-section in a given time.  The original field test of the device in a cotton picker was unsuccessful, primarily because of problems with stray light.  However, laboratory tests resulted in a high correlation (R² = 0.93) between the mass of cotton passing the device and the device’s output.  Cotton feed rate was reported to affect sensor performance, and airflow rate was a significant factor affecting sensor output. 

Thomasson et al. (1999) reported the design and fabrication of two experimental devices (device A and device B) for measuring the flow of pneumatically conveyed cotton.  Both devices worked on the principle of optical attenuation and consisted of a light-sensing bar and a light source in Device A, and an LED array and light-sensors in Device B.  In limited tests on a cotton picker in 1989, device A recorded data that were highly correlated with yield, with R² = 0.89 one day, and R² = 0.98 the next.  Differences are likely attributable to using sunlight as the light source in those early experiments.  When both devices were mounted in a gin-unloading duct, data from device B and actual flow rates were highly correlated, with an R² value of 0.90.  When the devices were mounted in a lint-cleaner-waste duct, data from device A and flow rates were highly correlated, with an R² value of 0.92. 

Zycom Corporation and Micro-Trak Systems, Inc., commercialized optical cotton yield monitors in 1997.  Both Micro-Trak and Zycom cotton yield monitors have been evaluated at the field level (Gvili, 1998; Durrence et. al. 1998).  These yield monitors have provided some useful data, but they have had some problems, one of which is that they are greatly affected by the buildup of dirt and dust on sensor surfaces; i.e., dirt and trash often introduce significant errors in the form of a drifting baseline.  Khalilian et al. (1999) developed an air-box, pressurized by the picker fan, to help keep the sensors clean.  The air-box completely encloses the sensor, effectively sealing it from environmental contamination.  This method was able to keep the sensor clean over several harvested loads.  In general, test results have shown that the commercial cotton yield monitors performed well when their sensor windows were clean and the systems were properly calibrated on a regular basis.  These conditions are very difficult to maintain in a commercial production situation. 

Thomasson and Sui (2000a) reported on an advanced cotton yield monitor with higher fundamental accuracy, insensitivity to dirt and dust buildup, and the ability to measure trash content in the seed cotton.  They later applied for a patent on the device (Thomasson and Sui, 2000b) to protect commercial licensing interests.  Sui et al. (2000) reported the development of a device, which simulates the pneumatic system on a cotton picker, for testing their yield monitor in the laboratory.  Finally, Sui and Thomasson (2001) reported on extensive field-testing of their yield monitor in 2000 and 2001 (Thomasson et. al., 2002).  Their tests showed a reliable system with average absolute errors between 5% and 6% in the 2000 harvesting season and between 3.7% and 4.9% in 2001 season. 

Current Research: 

Dr. J. A. Thomasson and Dr. R. Sui have successfully developed and extensively field-tested an optical cotton yield monitor.  The test results are very promising.  A patent on this cotton yield monitor is pending.  MSU has established an option agreement with a private company for the purpose of a future license for manufacturing and marketing this cotton yield monitor.  Some of the cotton yield-monitor technology, and much of the experience gained from developing it, can greatly benefit the development of yield monitors for other pneumatically conveyed crops.

References:

 Benjamin, C. E., M. P. Mailander, and R. R. Price.  2001. Sugar cane yield monitoring system.  ASAE Paper No. 01-1189.  St. Joseph, Mich.: ASAE. 

Durrence, J. S., C. D. Perry, G. Vellidis, D. L. Thomas, and C. K. Kvien.  1998.  Evaluation of commercially available cotton yield monitors in Georgia field conditions.  ASAE Paper No. 98-3106.  St. Joseph, Mich.: ASAE. 

Gvili, M.  1998.  Cotton yield sensor produces yield maps.  In Proc. Beltwide Cotton Conf., D. A. Richter, ed., pp. 1655-1657.  Memphis, Tenn.: National Cotton Council of Am. 

Khalilian, A., F. J. Wolak, R. B. Dodd, Y. J. Han.  1999.  Improved sensor mounting technology for cotton yield monitors.  ASAE Paper No. 991052.  St. Joseph, Mich.: ASAE. 

Sui, R. X., J. B. Wilkerson, W. E. Hart, and D. D. Howard.  1998.  Integration of neural network with a spectral reflectance sensor to detect nitrogen deficiency in cotton.  ASAE Paper No. 983104.  St. Joseph, Mich.: ASAE. 

Sui, R., J. A. Thomasson, and S. D. To.  2000.  Cotton-harvester-flow simulator for yield monitor development.  In Proc. 5th Int. Conf. on Precision Agric.  Bloomington, Minn.: Univ. of Minnesota Precision Agric. Center. 

Sui, R., and J. A. Thomasson.  2001.  Field testing of Mississippi State University cotton yield monitor.  In Proc. Beltwide Cotton Conf., P. Dugger and D. Richter, eds.  Memphis, Tenn.: Nat. Cotton Coucil Am. 

Thomasson, J. A., R. Sui, C. D. Perry, G. Vellidis, and G. Rains.  2002. Mississippi yield monitor: performance in farm testing.  In Proc. Beltwide Cotton Conf., P. Dugger and D. Richter, eds.  Memphis, Tenn.: Nat. Cotton Coucil Am (in press). 

Thomasson, J. A., D. A. Pennington, H. C. Pringle, E. P. Columbus, S. J. Thomson, and R. K. Byler.  1999.  Cotton mass flow measurement: experiments with two optical devices.  Applied Engineering in Agriculture 15(1):11-17.  

Thomasson, J. A., and R. Sui.  2000a.  Advanced optical cotton yield monitor.  In Proc. Beltwide Cotton Conf., P. Dugger and D. Richter, eds., 408-410.  Memphis, Tenn.: Nat. Cotton Coucil Am.

Thomasson, J. A., and R. Sui.  2000b.  Optical Mass Flow Sensor.  Provisional Patent Application.  Washington, D.C.: U.S. Patent Office. 

Vellidis, G., C. D. Perry, J. S. Durrence, D. L. Thomas, R. W. Hill, C. K. Kvien, T. K. Hamrita, G. Rains. 2001. The peanut yield monitoring system. Transactions of the ASAE.  Vol. 44(4): 775–785. 

Wilkerson, J. B., J. S. Kirby, W. E. Hart, and A. R. Womac.  1994. Real-time cotton flow sensor. ASAE Paper No. 94-1054.  St. Joseph, Mich.: ASAE.

Back to Top

Back to Menu