**6. Conclusions**

266 Remote Sensing of Biomass – Principles and Applications

The NDVI value, calculated from the LARS images collected for 65 days old crop, was suitable to estimate the total biomass (r2 = 0.760) and rice crop yield (r2 = 0.728). The protein content estimated through NDVI value was marginally suitable, capable to provide overall rice crop quality before-hand. The LARS system is suitable for real and near-real time crop parameter estimation, monitoring and evaluations. The NIR sensors can be substituted with any professional camera or cheap digital camera to optimize cost involved. The overall efficiency of the LARS system will be dependent on the sensors mounted on the helicopter. A skilled labor can easily handle the whole system with least supervision. The LARS not only replaces the satellite based image processing system but also ground level spectrophotometer, chlorophyll content measuring equipments. With little time, the system

Sugiura et al. (2007) mounted a thermal band camera on unmanned helicopter platform to estimate soil water status in paddy fields and correlation was obtained between the thermal image temperature and soil moisture content. The coefficient of determination (r2) for moisture content and temperature model at 10.00 a.m. and 3.00 p.m. were 0·69 and 0·64

Fig. 9. Soil moisture content estimation with LARS images (Sugiura et al., 2007)

The r2 between moisture content and difference in temperature was 0·42. The development was intended assisting in proper irrigation scheduling and monitoring water stressed situations for rain-fed cropping. Ishii et al. (2005) developed a system that can generate a map regarding crop status obtained by mounting an imaging sensor on an unmanned helicopter. They achieved an accuracy of 38 cm using RTK GPS receiver and GDS unit. The maps are accurate enough to be used for variable rate nutrients and pesticides application

Lenthe et al. (2007) used unmanned helicopter based IR thermography imaging system as a tool for monitoring the microclimatic conditions promoting incidence and severity of diseases within wheat fields with a high spatial resolution. Zhou et al. (2009) used R44 helicopter for aerial electrostatic spraying system. The results of the studies showed that electrostatic spraying with helicopters could produce uniform and fine droplets with better droplet adhesion and distribution, higher depositing efficiency, lower environmental

will be ready for taking images, for instance, just after rainfall.

**5. Field application of LARS systems** 

**4.4 Discussions** 

respectively (Figure 9).

for the farmland.

A radio-controlled helicopter-based LARS system can be used to acquire multispectral images over a rice canopy to estimate rice yield. The study indicated that the LARS platform could substitute for satellite-based and costly airborne remote sensing system. Images are obtained successfully by the multispectral camera mounted on the radio-controlled helicopter at a height of 20 m over rice plots. Rice yield and total biomass were found to be significantly different at the 0.05 and 0.1 significance levels, respectively, under different N treatment regimes. The relationship between NDVILARS and NDVISPECTRO (r2 = 0.897, RMSE = 0.012) shows the applicability of LARS sensor-based images for estimating NDVI values, which varied over the five levels of applied N. A linear regression model shows a good fit (r2 = 0.728, RMSE = 0.458 ton ha-1) for estimating total biomass for rice using LARS imagebased NDVI values. A linear model (r2 = 0.760, RMSE = 0.598 ton ha-1) indicates that rice yield could also be predicted with NDVI values derived from LARS images. The protein content can be positively estimated well in advance to actual crop harvesting. The regression model procedure outlined herein can be followed for larger rice fields by recording crop input rates and acquiring LARS images.
