Abstract
Considering that agricultural production is characterized by vast areas, scattered fields and long crop growth cycles, intelligent wireless sensor networks (WSNs) are suitable for monitoring crop growth information. Cost and coverage are the most key indexes for WSN applications. The differences in crop conditions are influenced by the spatial distribution of soil nutrients. If the nutrients are distributed evenly, the crop conditions are expected to be approximately uniform with little difference; on the contrary, there will be great differences in crop conditions. In accordance with the differences in the spatial distribution of soil information in farmland, fuzzy c-means clustering was applied to divide the farmland into several areas, where the soil fertility of each area is nearly uniform. Then the crop growth information in the area could be monitored with complete coverage by deploying a sensor node there, which could greatly decrease the deployed sensor nodes. Moreover, in order to accurately judge the optimal cluster number of fuzzy c-means clustering, a discriminant function for Normalized Intra-Cluster Coefficient of Variation (NICCV) was established. The sensitivity analysis indicates that NICCV is insensitive to the fuzzy weighting exponent, but it shows a strong sensitivity to the number of clusters.
Highlights
The real-time monitoring and exact diagnosis of crop growth are instructive to control crop development and enhance crop yields
normalized intra-cluster coefficient of variance (NICCV) can preferably recognize the data with different complexities in spatial distribution and the performance of the optimal cluster number determined by NICCV is superior to that of fuzziness performance index (FPI) and NCE
The sensitivity analysis indicated that NICCV is not sensitive to the cluster process parameter m of fuzzy c-means clustering when determining optimal classification number, it is found to be sensitive to FPI and NCE
Summary
The real-time monitoring and exact diagnosis of crop growth are instructive to control crop development and enhance crop yields. Aitsaadi et al [26] divided the target area into key and non-key monitoring areas to non-uniformly deploy the sensor nodes in a target area by using artificial potential fields and a tabu search heuristic algorithm, where the deployment density of nodes was connected with the probability of the key monitoring events in the target area This method was appropriate for application scenarios like fire monitoring, while in crop production and management, the monitoring probability density of crop growth information throughout the farmland was similar, it failed to satisfy the demands of providing crop growth information in a wide-area farmland environment, namely full coverage and low costs. The WSN deployed by this method was able to completely monitor the crop growth information under a wide-area farmland environment, while at the same time it could greatly decrease the number (and cost) of deployed sensor nodes
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
