Preharvest Monitoring of Biomass Production - Engineering and Science of Biomass Feedstock Production and Provision - page 67

real-time remote sensing system, (2) ground truth data collection vehicle, and (3)
UAV-based remote sensing system. These three different sensing systems and their
application to monitor energy crops such as switchgrass and Miscanthus are
described in the rest of the chapter [ 30 - 33 ].

4.7 S tand -A lone T ower -B ased R eal -T ime R emote S ensing

4.7.1 Tower Remote Sensing Principle and Instrumentations

To perform site-specifi c and seasonal monitoring of biomass energy crop growth
conditions, a stand-alone tower-based crop sensing system as indicated in Fig. 4.6
was designed and built on the Energy Farm of the University of Illinois at Urbana-
Champaign [ 34 ]. The tower is erected at the center of the fi eld and is equipped with
a motorized multispectral camera with lens controller for zoom and focus adjust-
ment and pan-tilt device and controller for horizontal (0°-355°) and vertical (0°-90°)
movement as indicated in Fig. 4.6a . The presets according to the fi eld distribution
were established using the caller identifi cations and automatic rotations of the
pan/tilt device that had been developed. The lens motorization was developed exter-
nally and used two motors to control zoom and focus. The tower-based system
captures near-real-time RGB and CIR images of four fi elds growing four different
crops, namely, Miscanthus, switchgrass, mixed prairie, and corn. The layout of the
fi eld is depicted in Fig. 4.6b . A Labview-based real-time algorithm was developed
to capture images from the fi eld over the growing seasons. Initially, 91 preset
positions were set to cover each of the fi elds. The 50-mm fi xed focal length was
chosen to capture images. The NIR, red, and green channels were averaged in the
image-acquisition process. The tower coordinates and the ground reference points
were surveyed using an RTK global position system (GPS) unit. The stand-alone
images for the reference points present the crop response and physiological changes.
Four different ground reference points for Miscanthus (M1, M2, M3, and M4) were
observed during the growing season and data was collected from early spring to
winter during 2009 and 2012.
The sensing system was established during August 2009, and images were ready
for acquisition from September 2009. The ground reference points for mixed prairie
grass (P1, P2, P3, and P4) were placed inside an 8-m by 8-m plot to track the vegeta-
tive responses. The reference points for switchgrass (S1, S2, S3, and S4) were
marked inside the fi eld. The fi eld spectrometry was limited to three energy grasses,
however, for geo-referencing the corner points for four fi elds (i.e., corn, Miscanthus,
switchgrass, and prairie grass were surveyed).
The stand-alone camera sensor system was developed with a four-band MS4100,
a multispectral charged couple device (CCD) camera (Geospatial), a pan/tilt device
(PT570P medium duty) and receiver (LRD41C21/22 Legacy), and a lens controller
(Fig. 4.7 ). The multispectral camera was a digital progressive scan camera with a
high resolution of 1,920 × 1,080 pixels. In contrast to a normal CCD camera, the