Search

Real-time monitoring of crop responses to environmental deviations represents a new avenue for applications of remote and proximal sensing. Combining the high-throughput devices with novel machine learning (ML) approaches shows promise in the monitoring of agricultural production. The 3 × 2 multispectral arrays with responses at 610 and 680 nm (red), 730 and 760 nm (red-edge) and 810 and 860 nm (infrared) spectra were used to assess the occurrence of leaf rolling (LR) in 545 experimental maize plots measured four times for calibration dataset (n = 2180) and 145 plots measured once for external validation. Multispectral reads were used to calculate 15 simple normalized vegetation indices. Four ML algorithms were assessed: single and multilayer perceptron (SLP and MLP), convolutional neural network (CNN) and support vector machines (SVM) in three validation procedures, which were stratified cross-validation, random subset validation and validation with external dataset. Leaf rolling occurrence caused visible changes in spectral responses and calculated vegetation indexes. All algorithms showed good performance metrics in stratified cross-validation (accuracy >80%). SLP was the least efficient in predictions with external datasets, while MLP, CNN and SVM showed comparable performance. Combining ML with multispectral sensing shows promise in transition towards agriculture based on data-driven decisions especially considering the novel Internet of Things (IoT) avenues. [ABSTRACT FROM AUTHOR]

Sensor networks, as a special subtype of wireless networks, consist of sets of wirelessly connected sensor nodes often placed in hard-to-reach environments. Therefore, it is expected that sensor nodes will not be powered from the power grid. Instead, sensor nodes have their own power sources, the replacement of which is often impractical and requires additional costs, so it is necessary to ensure minimum energy consumption. For that reason, the energy efficiency of wireless sensor networks used for monitoring environmental parameters is essential, especially in remote networking scenarios. In this paper, an overview of the latest research progress on wireless sensor networks based on LoRa was provided. Furthermore, analyses of energy consumption of sensor nodes used in agriculture to observe environmental parameters were conducted using the results of real measurements in the field, as well as simulations carried out based on collected data about real equipment. Optimization methods of energy consumption, in terms of choosing the appropriate data collection processes from the conducted field measurements, as well as the settings of network radio parameters imitating real conditions used in conducted simulations were highlighted. In the analyses, special emphasis was placed on choosing the optimal packet size. Unlike in other papers analyzing energy efficiency of LoRa communication, in this paper, it was proven that the adjustment of the transmission speed to the actual size of the packet is important for better energy efficiency of communication and that it can reduce energy consumption considerably. Moreover, in the paper, the contents of a packet that can be used in precision agriculture is suggested in order to prove that the 6-bit packet is sufficient for energy-efficient collection of parameters from the environment, in contrast to the 11-bit packets used in standard commercially available equipment. [ABSTRACT FROM AUTHOR]