Application of thermal imaging combined with machine learning for detecting the deterioration of the cassava root

Freshness is an important parameter that is indexed in the quality assessment of commercial cassava tubers. Cassava tubers that are not fresh have reduced starch content. Therefore, in this study, we aimed to develop a new approach to detect cassava root deterioration levels using thermal imaging with machine learning (ML). An underlying assumption was that nonfresh cassava roots may have fermentation inside that causes a difference in the inner temperature of the tuber. This creates the opportunity for the deterioration level to be measured using thermal imaging. The features (pixel intensity and temperature) that were extracted from the region of interest (ROI) in the form of tuber thermal images were analyzed with ML. Linear discriminant analysis (LDA), k-nearest neighbor (kNN), support vector machine (SVM), decision tree, and ensemble classifiers were applied to establish the optimal classification modeling algorithms. The highest accuracy model was developed from thermal images of cassava roots captured in a darkroom under a control temperature of 25 °C in the measurement chamber. The LDA, SVM, and ensemble classifiers gave the best overall performance for the discrimination of cassava root deterioration levels, with an accuracy of 86.7%. Interestingly, under uncontrolled environmental conditions, the combination of thermal imaging plus ML gave results that were of lower accuracy but still acceptable. Thus, our work revealed that thermal imaging coupled with ML was a promising method for the nondestructive evaluation of cassava root deterioration levels.

A New Deep Learning Framework for Imbalance Detection of a Rotating Shaft

Rotor unbalance is the most common cause of vibration in industrial machines. The unbalance can result in efficiency losses and decreased lifetime of bearings and other components, leading to system failure and significant safety risk. Many complex analytical techniques and specific classifiers algorithms have been developed to study rotor imbalance. The classifier algorithms, though simple to use, lack the flexibility to be used efficiently for both low and high numbers of classes. Therefore, a robust multiclass prediction algorithm is needed to efficiently classify the rotor imbalance problem during runtime and avoid the problem's escalation to failure. In this work, a new deep learning (DL) algorithm was developed for detecting the unbalance of a rotating shaft for both binary and multiclass identification. The model was developed by utilizing the depth and efficacy of ResNet and the feature extraction property of Convolutional Neural Network (CNN). The new algorithm outperforms both ResNet and CNN. Accelerometer data collected by a vibration sensor were used to train the algorithm. This time series data were preprocessed to extract important vibration signatures such as Fast Fourier Transform (FFT) and Short-Time Fourier Transform (STFT). STFT, being a feature-rich characteristic, performs better on our model. Two types of analyses were carried out: (i) balanced vs. unbalanced case detection (two output classes) and (ii) the level of unbalance detection (five output classes). The developed model gave a testing accuracy of 99.23% for the two-class classification and 95.15% for the multilevel unbalance classification. The results suggest that the proposed deep learning framework is robust for both binary and multiclass classification problems. This study provides a robust framework for detecting shaft unbalance of rotating machinery and can serve as a real-time fault detection mechanism in industrial applications.

Unlocking Insights in IoT-Based Patient Monitoring: Methods for Encompassing Large-Data Challenges

The remote monitoring of patients using the internet of things (IoT) is essential for ensuring continuous observation, improving healthcare, and decreasing the associated costs (i.e., reducing hospital admissions and emergency visits). There has been much emphasis on developing methods and approaches for remote patient monitoring using IoT. Most existing frameworks cover parts or sub-parts of the overall system but fail to provide a detailed and well-integrated model that covers different layers. The leverage of remote monitoring tools and their coupling with health services requires an architecture that handles data flow and enables significant interventions. This paper proposes a cloud-based patient monitoring model that enables IoT-generated data collection, storage, processing, and visualization. The system has three main parts: sensing (IoT-enabled data collection), network (processing functions and storage), and application (interface for health workers and caretakers). In order to handle the large IoT data, the sensing module employs filtering and variable sampling. This pre-processing helps reduce the data received from IoT devices and enables the observation of four times more patients compared to not using edge processing. We also discuss the flow of data and processing, thus enabling the deployment of data visualization services and intelligent applications.

Multi-class Classification of Agricultural Data Based on Random Forest and Feature Selection

Agricultural production and operation produce a large amount of data, which hides valuable knowledge. Data mining technology can effectively explore the connection between various factors from the massive agricultural data. Classification prediction is one of the most valuable agricultural data mining techniques. This paper presents a new algorithm consisting of machine learning algorithms, feature ranking method and instance filter, which aims to enhance the capability of the random forest algorithm and better solve the problem of agricultural multi-class classification. The performance of the new algorithm was tested by using four standard agricultural multi-class datasets, and the experimental results showed that the newly proposed method performed well on all datasets. Among them, substantial rise in classification accuracy is observed for Eucalyptus dataset. Applying random forest algorithm on Eucalyptus dataset results in classification accuracy as 53.4% and after applying the new algorithm (rough set) the classification accuracy significantly increases to 83.7%.

Energy Conservation for Internet of Things Tracking Applications Using Deep Reinforcement Learning

The Internet of Things (IoT)-based target tracking system is required for applications such as smart farm, smart factory, and smart city where many sensor devices are jointly connected to collect the moving target positions. Each sensor device continuously runs on battery-operated power, consuming energy while perceiving target information in a particular environment. To reduce sensor device energy consumption in real-time IoT tracking applications, many traditional methods such as clustering, information-driven, and other approaches have previously been utilized to select the best sensor. However, applying machine learning methods, particularly deep reinforcement learning (Deep RL), to address the problem of sensor selection in tracking applications is quite demanding because of the limited sensor node battery lifetime. In this study, we proposed a long short-term memory deep Q-network (DQN)-based Deep RL target tracking model to overcome the problem of energy consumption in IoT target applications. The proposed method is utilized to select the energy-efficient best sensor while tracking the target. The best sensor is defined by the minimum distance function (i.e., derived as the state), which leads to lower energy consumption. The simulation results show favorable features in terms of the best sensor selection and energy consumption.