**4.1.1 Calculations**

The rice biomass (threshed rice plant without the grain) of three sampled areas, 4 m2 each, were collected and weighted. The moisture content (w.b.) of the threshed rice plant was estimated using standard method. The dry weight of the threshed rice plant was estimated and converted into the total biomass weight per ha i.e. (ton/ha).

$$\text{Total biomass (ton/ha)} = \text{(100 - M.C.)} \times \text{Biomass} \text{Wt.} \times \frac{10000}{12 \times 1000} \tag{3}$$

Where,

Total biomass: Weight of rice plant (without rice grain) in ton/ha

BiomassWt: Weight of threshed rice plant (without rice grain)

M.C.: Moisture content of weighed rice plant (w.b.)

Total oven-dried (Abdullah et al., 1992) biomass was ranged from 3.58 to 7.36 ton ha-1 for the different treatments (Table 4). Total dry biomass weight between the treatments showed significant differences at the 0.10 level but no significant difference between replicates.


Table 3. Total biomass (ton ha-1) of the experimental plots

Linear calibrations curves were developed in SAS 9.1 to estimate the biomass from NDVI index values calculated from LARS images. From these results, NDVILARS could explain 76% of the variation in biomass weight (r2 = 0.760, RMSE = 0.598 ton ha-1, Figure 6).

Rice Crop Monitoring with Unmanned Helicopter Remote Sensing Images 265

The regression model, developed for rice yield with NDVI index value in SAS 9.1, indicated a good fit (r2 = 0.728, RMSE = 0.458 ton ha-1, Figure 7). Variation among the replicates might

0.85 0.88 0.91 0.94 0.97 1

Protein content is one of the major food nutrients to determine quality of the food-grain. It could be measured as the total available nitrogen content in the food stuff (Kennedy, 1995). The rice was powdered and sieved before testing for total nitrogen with standard method. The linear model of total nitrogen against NDVILARS (with r2 = 0.591, Figure 8) showed positive relationships, and would be useful to the farmers, as they can get idea of quality of rice grain well in advance, at booting stage (from the image taken during booting stage).

**NDVI LARS**

y = 22.753x - 18.342 <sup>2</sup> R = 0.7283 RMSE = 0.4581

be due to initial nutrient levels present in the soil from randomly selected plots.

Rep. 1 Rep. 2 Rep. 3

0.00

**4.3 Estimation of protein content** 

Fig. 7. Estimation of rice yield with NDVILARS values.

Fig. 8. Estimation of protein content with NDVILARS values.

0.75

1.50

2.25

3.00

**-1**

**Yield (tons ha )**

3.75

4.50

Fig. 6. Estimation of biomass with NDVILARS values

### **4.2 Estimation of rice yield**

The rice crop was harvested from three sample areas of 4 m2 from each plot, 102 days after sowing for this experiment. The moisture content (% w.b.) at the time of weighing was estimated using a field moisture meter (Kett PM600, Ohta-Ku, Tokyo, Japan). The yield of each plot (100 m2 area) was estimated as the average of three sampled areas and converted to a ton-per-hectare area using the following equation. Rice yield was estimated at 14% moisture content (MC) for each treatment (Field crop report, 1998).

$$Yield\\_\{ton\ ha^{-1}\} = \frac{(100 - MC) \times RV \times 10000}{86 \times A \times 1000} \tag{4}$$

Where,

MC = moisture content (% wet basis)

RW = weight of rice (kg)

A = harvested area (m2)

Rice yield, ranged from as low as 1.88 ton ha-1 (0 kg ha-1 N) to 3.68 ton ha-1 (132 kg ha-1 N) based on a 14% MC, illustrates the effectiveness of the fertilizer treatment rates on rice yield (Table 4). The crop yield variation was also tested for statistical significance (Johnson and Bhattacharyya, 2001). Yield data between the treatments showed significant differences at the 0.10 and 0.05 levels, whereas differences were not significant among the replicates


Table 4. Rice yield (ton ha-1) of the experimental plots

The regression model, developed for rice yield with NDVI index value in SAS 9.1, indicated a good fit (r2 = 0.728, RMSE = 0.458 ton ha-1, Figure 7). Variation among the replicates might be due to initial nutrient levels present in the soil from randomly selected plots.

Fig. 7. Estimation of rice yield with NDVILARS values.
