IAPRS & SIS, Vol.34, Part 7, “Resource and Environmental Monitoring”, Hyderabad, India, 2002 
REMOTE SENSING FOR PRECISION AGRICULTURE 
Lei Tian, Ph.D. 
Professor of Agricultural Engineering, 
Director of Illinois Laboratory for Agricultural Remote Sensing, University of Illinois at Urbana- 
  
Champaign, 1304 W. Pennsylvania Ave., Urbana, IL 61801, USA - lei-tianQ uiuc.edu 
Commission VI, WG VI/A 
KEY WORDS : Machine-vision, Precision Farming, Weed Mapping, 
[sninol] 
ABSTRACT: 
To demonstrate sensor-based precision farming concept, this 
with the commercial map-driven-ready sprayer, an 
applications. Remote sensing data was use 
The machine vision guided system was specia 
were used to cover the target area. To increase the delivery accuracy, 
o evaluate the effectiveness and performance under varying commercial field conditions. Using the on- 
ferenced chemical input maps (equivalent to weed maps) were also recorded in real-time. The maps 
integrated system was tested t 
board differential GPS, geo-re 
Herbicide Application, Remote Sensing. 
paper studied and compared the applications of the remote sensing system 
d a real-time machine vision guided "smart sprayer" in selective herbicide 
d to identify weed infestation area and simulate map-driven selective herbicide application. 
lly designed to work under outdoor variable lighting conditions. Multiple vision sensors 
each individual spray nozzle was controlled separately. The 
generated with this system have been compared with other sensing and referencing remote sensing systems. 
1. INTRODUCTION 
The potential of precision agriculture is limited by the lack of 
appropriate measurement and analysis techniques for 
agronomically important factors (National Research Council, 
1997). High accuracy sensing and data management tools must 
be developed and validated to support both research and 
production. The limitation in data quality/availability has become 
a major obstacle in the demonstration and adoption of the 
precision technologies (Stafford, 1996). The development of 
sensors lags behind other enabling technologies. Because of this 
limitation, much of the research on site-specific management 
(SSM) has involved only empirical comparisons of SSM versus 
uniform field management and has not quantitatively identified 
cause and effect relationships that determine the outcome. As a 
result, most comparisons have not detected significant benefits in 
either productivity or environmental quality despite theoretical 
predictions that such benefits should occur. 
In some field operations, precision farming concept is very hard 
to realize. For example, the site-specific pest control is 
considered to be one of most difficult tasks. The practice of 
blanket application of herbicides results in over-application in 
areas of low or no weed infestation causing environmental 
problems such as soil and groundwater pollution. If herbicides 
could be efficiently applied in a spatially varying manner based 
on weed type and population, less herbicide would be used and 
the herbicide runoff and leaching would be drastically reduced. 
The benefits of site-specific application of herbicides are larger as 
the weed population and level of patchiness of weeds increase in 
a field (Oriade et al., 1996). Site-specific application of herbicide 
based on a weed map could save 4046, to 60% of herbicide 
(Brown and Steckler, 1995; Tian, et al, 1999). Two types of 
variable-rate herbicide application systems are currently under 
investigation, namely map-based systems and real-time sensor- 
based systems. If accurately done, a map-based system could be 
less expensive, and simple to use as compared to a real-time 
system. In a real-time system, data collection, decision-making 
and implementation of the management decision are done at the 
same time in one pass of the machine. In a near-real-time system, 
the management decision is made and implemented within a few 
hours or a day after data collection. The major concern in a map- 
based-system is *he quality of the maps used for precision 
application. Conventional methods of weed mapping based on 
field scouting systems (Heisel et al., 1996; Stafford et al., 1996; 
Krueger et al. 1998) are time consuming, laborious, and 
expensive though it is very accurate, and has the capability to 
identify weed distribution based on species, population, and age. 
Vehicle-mounted weed mapping systems (Rew et al., 1996) are 
slow and expensive as compared to remote sensing systems. On 
the other hand, real-time systems for site-specific herbicide 
application (Brivot and Marchant, 1996; Tian et al., 1999) are 
more accurate and efficient in site-specific weed control, but may 
be expensive for many major crops at present time. A few real- 
time field systems have been developed. The photosensor-based 
plant detection systems (Shearer and Jones, 1991; Hanks, 1996) 
can detect all the green plants (weed and crop plants) and spray 
only on the plants. For high-value crops, high-accuracy machine 
vision and control systems have been studied for outdoor field 
applications in California (Tian et. al., 1997, Lee and Slaughter, 
1998). 
2. METHODS 
Both remote sensing and ground based sensing system were 
employed in the experiment of evaluating the sensing systems for 
precision chemical application systems. To do this, we used a 
near-real-time aerial imaging system to collect color near infrared 
images of the experiment fields. 
    
2.1 Ael 
The aei 
GPS re 
infrarec 
NY) w 
CCD s 
resoluti 
range « 
Eastma 
sensor 
(500 ni 
810 nn 
mm to 
(Trimb 
camera 
was ta 
proces: 
mounte 
acquire 
2.2 Sp 
The sn 
in Fig 
system 
applica 
control 
The lai 
Tyler.‘ 
driven 
differe 
positio 
connec 
nozzle: 
were C 
Figure 
time aj