The rise of Internet of Things (IoT) technology has moved the digital world in a new direction and is considered the third wave of the information industry. To meet the current growing demand for food, the agricultural industry should adopt updated technologies and smart agriculture based on the IoT which will strongly enable farmers to reduce waste and increase productivity. This research presents a novel system for the application of IoT technology in agricultural soil measurements, which consists of multiple sensors (temperature and moisture), a micro-processor, a microcomputer, a cloud platform, and a mobile phone application. The wireless sensors can collect and transmit soil information in real time with a high speed, while the mobile phone app uses the cloud platform as a monitoring center. A low power consumption is specified in the hardware and software, and a modular power supply and time-saving algorithm are adopted to improve the energy effectiveness of the nodes. Meanwhile, a novel soil information prediction strategy was explored based on the deep Q network (DQN) reinforcement learning algorithm. Following the weighted combination of a bidirectional long short-term memory, online sequential extreme learning machine, and parallel extreme machine learning, the DQN Bi-OS-P prediction model was obtained. The proposed data acquisition system achieved a long-term stable and reliable collection of time-series soil data with equal intervals and provided an accurate dataset for the precise diagnosis of soil information. The RMSE, MAE, and MAPE of the DQN Bi-OS-P were all reduced, and the R2 was improved by 0.1% when compared to other methods. This research successfully implemented the smart soil system and experimentally showed that the time error between the value displayed on the mobile phone app and its exact acquisition moment was no more than 3 s, proving that mobile applications can be effectively used for the real-time monitoring of soil quality and conditions in wireless multi-sensing based on the Internet of Things.
In response to the need for deploying the YOLOv4-Tiny model on resource-constrained Field-Programmable Gate Array (FPGA) platforms for rapid inference, this study proposes a general optimization acceleration strategy and method aimed at achieving fast inference for object detection networks. This approach centers on the synergistic effect of several key strategies: a refined resource management strategy that dynamically adjusts FPGA hardware resource allocation based on the network architecture; a dynamic dual-buffering strategy that maximizes the parallelism of data computation and transmission; an interface access latency pre-configuration strategy that effectively improves data throughput; and quantization operations for dynamic bit width tuning of model parameters and cached variables. Experimental results on the ZYNQ7020 platform demonstrate that this accelerator operates at a frequency of 200 MHz, achieving an average computing performance of 36.97 Giga Operations Per Second (GOPS) with an energy efficiency of 8.82 Giga Operations Per Second per Watt (GOPS/W). Testing with a metal surface defect dataset maintains an accuracy of approximately 90% per image, while reducing the inference delay per frame to 185 ms, representing a 52.2% improvement in inference speed. Compared to other FPGA accelerator designs, the accelerator design strategies and methods proposed in this study showcase significant enhancements in average computing performance, energy efficiency, and inference latency.
Over the last decade, various deep neural network models have achieved great success in image recognition and classification tasks. The vast majority of high-performing deep neural network models have a huge number of parameters and often require sacrificing performance and accuracy when they are deployed on mobile devices with limited area and power consumption. To address this problem, we present an SSD-MobileNet-v1 acceleration method based on network compression and subgraph fusion for Field-Programmable Gate Arrays (FPGAs). Firstly, a regularized pruning algorithm based on sensitivity analysis and Filter Pruning via Geometric Median (FPGM) was proposed. Secondly, the Quantize Aware Training (QAT)-based network full quantization algorithm was designed. Finally, a strategy for computing subgraph fusion is proposed for FPGAs to achieve continuous scheduling of Programmable Logic (PL) operators. The experimental results show that using the proposed acceleration strategy can reduce the number of model parameters by a factor of 11 and increase the inference speed on the FPGA platform by a factor of 9–10. The acceleration algorithm is applicable to various mobile edge devices and can be applied to the real-time monitoring of forest fires to improve the intelligence of forest fire detection.
Leaf moisture content (LMC) directly affects the life activities of plants and becomes a key factor to evaluate the growth status of plants. To explore a low-cost, real-time, rapid, and accurate method for LMC detection, this paper employs Ultra-High-Frequency Radio-Frequency Identification (UHF RFID) sensor technology. By reading the tag information attached to the back of leaves, the parameters of the RSSI, phase, and reading distance of the tags are collected. In this paper, we propose an enhanced Multi-Feature Fusion algorithm based on Hyperdimensional Computing (HDC) called MFFHDC. In our proposed method, the real-valued features are encoded into hypervectors and then combined with Multi-Linear Discriminant Analysis (MLDA) for the feature fusion of different features. Finally, a retraining method based on Cosine Annealing with Warm Restarts (CAWR) is proposed to improve the model and further enhance its accuracy. Tests conducted in the experimental forest show that the proposed mechanism can effectively predict the LMC. The model’s Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Coefficient of Determination (R2) reached 0.0195, 0.0255, and 0.9131, respectively. Additionally, comparisons with other methods demonstrate that the presented system performs excellently in most aspects. As a lightweight model, this study shows great practical application value, particularly for the limited data volume and low hardware costs.
Trunk water has an important influence on the metabolism and ecological balance of living trees, which affects the vegetation growth and moisture cycle of the whole forest ecosystem. The accurate and real-time measurement of moisture content (MC) is of vital guiding meaning to living tree cultivation and forest management. In this paper, a water content diagnosis system based on a wireless acoustic emission sensor network (WASN) was designed and implemented with the aim of the nondestructive detection of water content in living wood trunks. Firstly, the acoustic emission (AE) signal of the trunk epidermis was sampled at high speed; then, its characteristic parameters were calculated and transmitted wirelessly to the gateway. Furthermore, the optimal characteristic wavelet sequence was decomposed by the adaptive chirp mode decomposition (ACMD), and the improved grey wolf optimizer (IGWO) optimization XGBoost established the MC prediction model, which was improved by the multi-strategy joint optimization. Finally, field monitoring was carried out on Robinia Pseudoacacia, Photinia serrulata, Pinus massoniana and Toona sinensis. The average diagnostic accuracy reached 96.75%, which shows that the diagnosis system has excellent applicability in different working conditions.
Water plays an important role in various physiological activities of living trees. Measuring trunk moisture content (MC) in real-time without damage has important guiding significance for transpiration research in forest ecosystems. However, existing standing tree MC detection methods are either too cumbersome to install or cause different degrees of damage. Here, we propose a novel Internet of Things (IoT) monitoring system that includes wireless acoustic emission sensor nodes (WASNs) and underground soil MC sensor nodes to efficiently detect and diagnose the MC level of living tree trunks. After the characteristic parameters were collected by the two sensors, a feature selection and multi-sensory global fusion method for MC diagnosis was designed and developed and several statistical parameters were selected as the input variables to predict the heartwood MC level with a support vector machine (SVM) model. Moreover, to achieve the highest prediction accuracy, an improved sparrow search algorithm (ISSA) is applied to ensure the most suitable parameter combinations in a two-objective optimization model. Extensive experiments result in a fusion of the environment, and AE signals show that the proposed mechanism has better diagnostic performance than state-of-the-art methods and is more adaptable to the fluctuation of working conditions.