Sensor Placement Algorithm With Range Constraints for Precision Agriculture Figure 8. Sensor Placement for 40 Nodes (20-40). Figure 6. Sensor placement using SPARC for 30 Nodes. ESTIMATION OF MOISTURE LEVEL USING SPARC In this section, we describe the estimation of moisture level across the farm using SPARC algorithm for illustrating our approach. While the proposed method can be used to estimate other variables such as temperature, NDVI or others as well. We selected moisture due to the high variability. We obtained the moisture levels across the farm from RS data, then the moisture sensor measured the data across different plants as a change in resistance. In addition, to the average value, we obtained the variance across the farm from sensor readings. The moisture field was developed using the sensor field reconstruction approach proposed in this paper. Outline of the farm is shown in Figure 5, the problem is to deploy minimum sensors to estimate the variables of interest. We consider moisture level to be measured across the 80 trees in the farm and due to cost considerations only about 40 WSN nodes can be deployed. In addition, there are range constraints of the WSN due to the limited transmission capability of NRF24L01. We use the SPARC to deploy the sensors wherein range constraints are considered. The results of 30 and 40 sensor nodes deployment are shown in Figures 6 and 8, respectively. The nodes having crosses are the sensor placements with range constraints obtained from SPARC. The variations in error with the number of sensors with a maximum number of 30 WSN nodes with and without Figure 7. Figure 9. Variations in Cumulative Error with SPARC (20-30). Variations in Cumulative Error with SPARC (20-40). 12 IEEE A&E SYSTEMS MAGAZINE JUNE 2019