Manual Monitoring Research Articles

IoT-based implementation of rickshaws for real-time monitoring and measuring the weight of cattle

In the era of modern agriculture that is increasingly dependent on technology, livestock management has become crucial to increasing efficiency and productivity. An important aspect in livestock management is providing appropriate feed to fattening cattle. Manual monitoring of feed weight is often complex and prone to errors, which can have a significant impact on operational efficiency and result in losses. Accuracy in monitoring feed weight is the key to maintaining optimal health and growth of cattle. Internet of Things (IoT) technology is emerging as an innovative solution to overcome these challenges. The use of Angkong load cells, a tool connected to IoT, allows automatic monitoring of feed weight with a high level of precision. The test results show an error rate close to zero, with a Mean Absolute Percentage Error (MAPE) of around 0.158%, making the Angkong load cell a reliable tool. With this capability, farmers can monitor cow feed weight in real-time with minimal error rates. This not only increases the operational efficiency of the farm but also optimizes the health and growth of livestock more efficiently, having a positive impact on overall farm productivity. The aim of this research is to monitor the amount of feed given to cows with an adequate level of accuracy. Rickshaw load cells can be well suited for this use due to their ability to handle relatively large weights with fairly good accuracy, but do not necessarily have the level of precision required in laboratory measurements or the pharmaceutical industry.

Design Warehouse Information System to Increase Stock Accuracy

In its development, the warehouse is currently not only used as a room for storage but has been integrated with the information system used in carrying out its warehousing activities. One of the objectives of the warehousing information system is to achieve stock accuracy according to the Company's target. PT Zoomlion Indonesia Heavy Industry (PT. ZIHI) is one of the heavy equipment industry companies engaged in the provision of heavy equipment, construction and plantations that have used the SAP system in managing their inventory but have not reached the stock accuracy target expected by the Company because in SAP there is no warehouse module. The purpose of this study is to find out what causative factors can affect stock inaccuracies in Jakarta warehouses, find out the dominant factors that cause stock inaccuracies and to find out the improvements made in increasing stock accuracy in PT Zoomlion Indonesia's Jakarta warehouse. The methods used are the DEMATEL (Decision Making Trial and Evaluation Laboratory) method, five why analysis, fishbone diagrams, for further design of the bin location system database. The results showed that (1) the factors causing stock accuracy inaccuracies at PT Zoomlion Indonesia Heavy Industry were due to increased receiving, storage, and shipping activities and limited manpower, and were not supported by a system that facilitates the location of parts. Meanwhile, according to fishbone analysis, the factor diagram that causes stock accuracy is not optimal is done through the five whys approach, which consists of human factors, methods, materials and the environment, (2) the dominant factors that cause inaccuracies in stock accuracy occur at risk (E10) There is no SAP support system for registering Bin location, (3)The web database stock location design developed is a form of improvement effort from the manual monitoring system that is not running smoothly. This stock location database will provide information about materials while in the warehouse room. Therefore, the use of a mobile- based web server can make it easier for users to access the system from a smartphone device.

Traffic flow detection method based on improved SSD algorithm for intelligent transportation system.

With the development of the new generation communication system in China, the application of intelligent transportation system is more extensive, which brings higher demands for vehicle flow detection and monitoring. Traditional traffic flow detection modes often cannot meet the high statistical accuracy requirement and high-speed detection simultaneously. Therefore, an improved Inception module is integrated into the single shot multi box detector algorithm. An intelligent vehicle flow detection model is constructed based on the improved single shot multi box detector algorithm. According to the findings, the convergence speed of the improved algorithm was the fastest. When the test sample was the entire test set, the accuracy and precision values of the improved method were 93.6% and 96.0%, respectively, which were higher than all comparison target detection algorithms. The experimental results of traffic flow statistics showed that the model had the highest statistical accuracy, which converged during the training phase. During the testing phase, except for manual statistics, all methods had the lowest statistical accuracy on motorcycles. The average accuracy and precision of the designed model for various types of images were 96.9% and 96.8%, respectively. The calculation speed of this intelligent model was not significantly improved compared to the other two intelligent models, but it was significantly higher than manual monitoring methods. Two experimental data demonstrate that the intelligent vehicle flow detection model designed in this study has higher detection accuracy. The calculation speed has no significant difference compared with the traditional method, which is helpful to the traffic flow management in intelligent transportation system.

Enhanced Machine Learning-based Diagnostic Model for Lassa Fever

The research presents a machine learning-based model for the diagnosis of Lassa fever. The development of the ML Model involves the collection of the Lassa fever dataset from the Infectious Diseases Control Centre of the Federal Medical Centre, Owo Ondo state, it was preprocessed and Five machine learning algorithms namely Naïve Bayes (NB), Artificial Neural Network (ANN), Support Vector Machine (SVM), k-Nearest Neighbour (k-NN) and Decision Tree were trained and tested using 10-fold cross-validation technique and evaluated using standard metrics. Four of the classifiers had an accuracy of 90% and above while k-Nearest Neighbour had the lowest accuracy of 87.7%. Naïve Bayes and Support Vector Machine performed well with an accuracy of 93.2% and Precision of 97.0%. The AUC measure for the NB and SVM algorithm was calculated to be 92.9 %. The Diagnostic Accuracy (DA) of the Existing Method (EM) calculated using standard metrics and evaluated through simple statistical techniques was 61.1%. The Machine Learning (ML) models performed better and more effectively in aiding the diagnosis of Lassa fever cases than the EM as determined from the dataset. The recommended ML model also has the potential to improve Medical diagnosis and eliminate the cumbersome, time-consuming and costly nature of manual monitoring of Lassa suspects in the current practice and can serve as a basis for further investigation/implementation of ML applications in the diagnosis of other similar diseases.

FruitScan Pro : Fruit Disease and Quality Detection Application

This research paper proposes a revolutionary approach to agriculture by combining the latest image processing technologies with machine learning. The agricultural industry, which is vital for India’s economy, needs efficient methods for monitoring fruit health. Manual monitoring is time consuming and often inadequate. In this paper, we propose a smart farming approach using image processing techniques. This approach improves image quality, segmenting images, extracting features, and classifying diseases and their cures. Fruit diseases are a major economic challenge around the world. Therefore, it is necessary to identify them quickly. Our proposed system offers fast and automatic disease detection. Preprocessing, extracting features, classifying, training/testing. The results prove its effectiveness for accurate disease detection and solutions. As the demand for agricultural productivity increases, this method will be a valuable tool for increasing fruit production with minimum human intervention. Image processing for effective implementation. Key Words: Fruit Disease, Solution, Image Preprocessing, Feature Extraction, Classification.

Real-Time Web-Based Monitoring System for Temperature, Humidity, and Solar Panels in Ramie Drying Facilities

Purpose: To address the manual monitoring challenges in processing ramie fibers, especially during drying. The purpose is to create a monitoring system that oversees room temperature, humidity, and the status of solar panels, crucial factors in ramie productivity.Methods: Real-time web-based system development that monitors room temperature, humidity, and the performance of solar panels in a ramie drying room using the Internet of Things architecture ESP32 with communication through GSM SIM 800L in rural areas.Results: The system can display real-time information such as temperature data, humidity, and electrical energy parameters derived from the solar panel's utilization in the ramie drying room. By doing so, users gain efficiency and effectiveness in obtaining information, significantly enhancing ramie fiber productivity.Novelty: Integration of sensor instruments, low-power ESP32 microcontrollers, GSM Telecommunication, Solar Cell Energy as a power source, and a real-time web-based Monitoring Information System implemented in a ramie drying dome. This simplifies the monitoring process and optimizes limited resources such as space, energy, telecommunications, and human resources, which are typically constrained infrastructure in the ramie fiber agricultural system.

Anomaly Detection in Heterogeneous Data Using Graph Convolutional Networks

Closed-circuit television (CCTV) surveillance systems are ubiquitous tools for ensuring public and private safety. However, manual monitoring of surveillance footage is arduous and time-consuming, often resulting in delayed identification and response to potential threats. In this paper, we present a novel real-time threat detection system tailored for CCTV surveillance, employing cutting-edge deep learning methodologies, particularly graph convolutional networks (GCNs), to detect and classify high-movement levels within video frames. By treating videos as segments and delineating between anomalous (potentially threatening) and normal (safe) segments, our system facilitates continuous real-time monitoring of surveillance footage, enabling the timely identification of potential threats such as abuse, burglaries, explosions, shootings, fighting, shoplifting, road accidents, arson, robbery, stealing, assault, and vandalism. To assess the efficacy of our system, we conducted extensive experiments on a large-scale dataset comprising CCTV footage, yielding promising results. Additionally, we introduce a novel, comprehensive dataset comprising 128 hours of real-world surveillance videos, encompassing various anomalies and normal activities. This dataset serves dual purposes: general anomaly detection, and the recognition of specific anomalous activities. Our experimental findings demonstrate the superior performance of our method in anomaly detection compared to state-of-the-art approaches. Furthermore, we provide an analysis of several recent deep learning baselines for anomalous activity recognition, revealing the challenging nature of our dataset and the potential it holds for future research endeavors. Through the integration of advanced deep learning techniques and comprehensive dataset creation, our proposed system stands to significantly enhance the efficiency and effectiveness of CCTV surveillance, facilitating faster response times and bolstering security for individuals in diverse environments.

Unsupervised Conditional Diffusion Models in Video Anomaly Detection for Monitoring Dust Pollution.

Video surveillance is widely used in monitoring environmental pollution, particularly harmful dust. Currently, manual video monitoring remains the predominant method for analyzing potential pollution, which is inefficient and prone to errors. In this paper, we introduce a new unsupervised method based on latent diffusion models. Specifically, we propose a spatio-temporal network structure, which better integrates the spatial and temporal features of videos. Our conditional guidance mechanism samples frames of input videos to guide high-quality generation and obtains frame-level anomaly scores, comparing generated videos with original ones. We also propose an efficient compression strategy to reduce computational costs, allowing the model to perform in a latent space. The superiority of our method was demonstrated by numerical experiments in three public benchmarks and practical application analysis in coal mining over previous SOTA methods with better AUC, of at most over 3%. Our method accurately detects abnormal patterns in multiple challenging environmental monitoring scenarios, illustrating the potential application possibilities in the environmental protection domain and beyond.

An Improved Safety Belt Detection Algorithm for High-Altitude Work Based on YOLOv8

High-altitude work poses significant safety risks, and wearing safety belts is crucial to prevent falls and ensure worker safety. However, manual monitoring of safety belt usage is time consuming and prone to errors. In this paper, we propose an improved high-altitude safety belt detection algorithm based on the YOLOv8 model to address these challenges. Our paper introduces several improvements to enhance its performance in detecting safety belts. First, to enhance the feature extraction capability, we introduce a BiFormer attention mechanism. Moreover, we used a lightweight upsampling operator instead of the original upsampling layer to better preserve and recover detailed information without adding an excessive computational burden. Meanwhile, Slim-neck was introduced into the neck layer. Additionally, extra auxiliary training heads were incorporated into the head layer to enhance the detection capability. Lastly, to optimize the prediction of bounding box position and size, we replaced the original loss function with MPDIOU. We evaluated our algorithm using a dataset collected from high-altitude work scenarios and demonstrated its effectiveness in detecting safety belts with high accuracy. Compared to the original YOLOv8 model, the improved model achieves P (precision), R (recall), and mAP (mean average precision) values of 98%, 91.4%, and 97.3%, respectively. These values represent an improvement of 5.1%, 0.5%, and 1.2%, respectively, compared to the original model. The proposed algorithm has the potential to improve workplace safety and reduce the risk of accidents in high-altitude work environments.

A survey of smart dustbin systems using the IoT and deep learning

With massive population growth and a shift in the urban culture in smart cities, the constant generation of waste continues to create unsanitary living conditions for city dwellers. Overflowing solid waste in the garbage and the rapid generation of non-degradable solid waste produce a slew of infectious illnesses that proliferate throughout the ecosystem. Conventional solid waste management systems have proved to be increasingly harmful in densely populated areas like smart cities. Also, such systems require real-time manual monitoring of garbage, high labor costs, and constant maintenance. Monitoring waste management on a timely basis and reducing labor costs is scarcely possible, realistically, for a municipal corporation. A Smart Dustbin System (SDS) is proposed that is to be implemented in densely populated urban areas to ensure hygiene. This paper undertakes a comprehensive analysis of the application of smart dustbin systems, following an extensive literature review and a discussion of recent research that is expected to help improve waste management systems. A current SDS used in real-time is implemented with the most recent advances from deep learning, computer vision, and the Internet of Things. The smart dustbin system used in day-to-day life minimizes the overloading of bins, lowers labor costs, and saves energy and time. It also helps keep cities clean, lowering the risk of disease transmission. The primary users of the SDS are universities, malls, and high-rise buildings. The evolution of the SDS over the years with various features and technologies is well analyzed. The datasets used for Smart Waste Management and benchmark garbage image datasets are presented under AI perception. The results of the existing works are compared to highlight the potential limitations of these works.

IoT Based Onion Storage Monitoring and Automation Control System

Onions are a staple crop globally, consumed in various culinary dishes and food products. However, their post-harvest preservation presents a significant challenge. Onions are sensitive to moisture, temperature, and humidity variations, making them prone to rot and spoilage during storage. Traditional storage methods often rely on periodic manual checks and manual intervention, which can be ineffective and resource-intensive. This project aims to address these challenges by implementing an Internet of Things (IoT) based solution that leverages moisture sensors, temperature and humidity sensors, microcontrollers, and an IoT platform to continuously monitor and regulate storage conditions, ensuring optimal onion quality and extended shelf life. Onions, a fundamental and widely consumed vegetable in global culinary traditions, face significant post- harvest preservation challenges. Sensitivity to variations in storage conditions, such as temperature, humidity, and moisture levels, often leads to spoilage and a rapid decline in quality. Traditional storage methods reliant on manual monitoring prove resource-intensive, prone to human error, and lack real-time insights into the evolving storage environment. This project introduces an innovative solution, an Internet of Things (IoT)-based onion storage monitoring system, to address these challenges. The system deploys sensors to continuously collect data on storage conditions, transmitting it to a cloud-based platform for real-time analysis. The analysis identifies trends and anomalies, triggering alerts and recommendations via a user-friendly interface. The anticipated outcomes of this project include enhanced onion quality, reduced spoilage, and increased resource efficiency for farmers and storage facility operators. This technology-driven approach contributes to the availability of high-quality onions in the market, serving both agricultural and culinary sectors. The IoT-based onion storage monitoring system represents a substantial step forward in post-harvest preservation, marking a promising advancement in the agriculture industry

Feasibility of Leveraging Consumer Wearable Devices with Data Platform Integration for Patient Vital Monitoring in Low-Resource Settings.

Manual monitoring of vital signs, which often fails to capture the onset of deterioration, is the main monitoring modality in most Ghanaian hospitals due to the high cost and inadequate supply of patient bedside monitors. Consumer wearable devices (CWDs) are emerging, relatively low-cost technologies for continuous monitoring of physiological status; however, their validity has not been established in low-resource clinical settings. We aimed to (1) investigate the validity of the heart rate (HR) and oxygen saturation (SpO2) data from two widely used CWDs, the Fitbit Versa 2 and Xiaomi Mi Smart Band 6, against gold standard bedside monitors in one Ghanaian hospital and (2) develop a web application to capture and display CWD data in a clinician-friendly way. A healthy volunteer simultaneously wore both CWDs and blood pressure cuffs to measure HR and SpO2. To test for concordance, we conducted the Bland-Altman and mean absolute percentage error analyses. We also developed a web application that retrieves and displays CWD data in near real time as text and graphical trends. Compared to gold standards (patient monitor and manual), the Fitbit Versa 2 had 96.87% and 96.67% measurement accuracies for HR, and the Xiaomi Mi Smart Band 6 had 94.24% and 93.21% measurement accuracies for HR. The Xiaomi Mi Smart Band 6 had 98.79% measurement accuracy for SpO2. The strong concordance between CWD and gold standards supports the potential implementation of these devices as a novel method of vital sign monitoring to replace manual monitoring, thus saving costs and improving patient outcomes. Further studies are needed for confirmation.

Design and Implementation of an IoT-based automated EC and pH Control System in an NFT-based Hydroponic Farm

In hydroponics, the growth of plants is primarily dependent on the precise configuration of the nutrient solution supplied to them. Continuous monitoring and control of the nutrients in the hydroponic tank is essential to maintain optimal growth requirements at all times. Manual monitoring and control of the system is inefficient, as it requires human intervention and can lead to errors and variations in the nutrients supplied. This study presents a system to automatically track pH, current nutrient, and water level in the tank and supply the exact amount of pH solution, nutrient solution, and water whenever needed. The proposed system, when integrated with IoT technology, can help achieve better control over the farm and higher productivity with a reduced amount of manual labor.

Two-stage helmet security detection model combining CNN and MTCNN algorithm

Physical safety is of utmost importance for workers in an industrial and construction environment. Fatalities related to the violation of wearing a safety helmet account for a large proportion of occupational deaths. Hence, the requirement of a high-efficiency automatic system to monitor and promote wearing helmets is of vital importance, reducing hours from manual monitoring, facilitating efficient education and enforcement campaigns that increase industrial safety. In this paper, a combined Convolutional Neural Network (CNN) algorithm and Multi-Task CNN (MTCNN) to monitor violations of safety helmets is proposed. All images are sourced from industrial and construction locations, undergoing various image processing techniques before model detection. The proposed model includes the MTCNN locating individuals' heads from the picture data, followed by a CNN assessment to determine helmet wearing. Under the same evaluation criterion, the result, where an obvious detection accuracy improvement is shown, illustrates the effectiveness of the proposed model in the comparative study. Therefore, the research demonstrates the effectiveness of incorporating multi-scale feature extraction techniques for enhancing model performance. This establishes the groundwork for future research within this domain, facilitating the creation of more advanced helmet detection or other relevant items recognition systems that can be utilized in real-world applications.

IOT Based Smart Solution for Preventing Fall Injuries at Workplace

In India fall injuries are increasing rapidly at workplace not only in the industrial sector even though construction and private sector too. Fall injury is the highest cause of death/ fatal/severe injuries. The severity of working at height incident depending upon the nature of job, height of application, uneven/cluttered surface etc.In this research paper a unique solution has been studied to bring revolutionary change in personal protective equipment which is used while working at height. There are many technologies are available, in chemical industry if any parameter deviate automatically machine will be shut-down and safe operating protocols will be activated but in conventional personal protective equipment it is hardly say there is still area of research where the technological improvement is needed to bring the robust safety system. Air bags are a very good example to correlate the smart PPE, if any passenger vehicle have air bag system collision with other vehicle automatically within milli second the airbag will be activated which saves the life of driver and passenger as well. Similarly Smart PPE has unique provision if PPE compliance is deviated at workplace immediately the violation shall be captured, and warning alert shall be given to user to remind them to compliance. This will help us to improve workplace safety. It is an error proof system that neither required manual surveillance nor manual recording of violations. In this research paper IOT internet of thing based smart safety harness has been studied and the workplace deviation were compared earlier with manual monitoring data to analyze the human error with respect to compliance of personal protective equipment

Long-term field studies of a distributed network of sensors for environmental radiological monitoring

The W-MON project goal is to establish an automatic control mechanism of the presence of radioactive material in conventional waste containers at CERN using a distributed network of interconnected low-power radiation sensors. This network facilitates continuous data recording, transfer and storage in a database while allowing online and offline data analysis, in addition to alarm triggering. Data transmission, processing and evaluation is achieved by a centralized IoT end-to-end data architecture that has been developed for real-time monitoring and visualization of the radiation levels in waste containers. In this paper the results of field tests of the W-MON system described in two previous papers are presented for three different types of sensors. Estimation of failure detection probability, long-term stability tests and sensitivity studies carried out using radioactive samples of various activities placed in standard waste containers are described. A comparison between the manual monitoring procedure currently used at CERN and the W-MON system is discussed in detail.

Can 3D radiological calculations predict operational difficulties for rectal cancer?: A single center retrospective analysis.

Low anterior resection, performing total mesorectal excision with appropriate pelvic dissection to prevent local recurrence, is probably the most challenging type of surgery in colorectal surgery, especially in a narrow pelvis. In this study, we aimed to predict the operation difficulty of rectal cancer by comparing the operation time with 2D and 3D pelvimetry. Sixty-six patients who underwent total mesorectal excision after neoadjuvant chemoradiotherapy due to primary rectal cancer located in the middle and lower rectum (10 cm from the anus) were included in the study. Surgery notes were reviewed and data on demographic factors, tumor stage, duration of surgery, and types of surgery were collected, as well as pelvimetric parameters. All protocols had 2D T2-weighted sequences in 3 planes (axial, sagittal, and coronal). Pelvimetric measurements were made by measuring 8 pelvic lengths and 2 angles. Pelvis and tumor volume were measured by manual margin monitoring. In each slice, both pelvis and tumor boundaries were manually drawn individually in the sagittal plane. Pelvis and tumor volumes were calculated from the set of adjacent images by summing slice thickness and products of area measurements within the pelvis and tumor boundaries. In our results, no correlation was observed with operation time, including pelvic volume. Exception for this were interacetabular distance and tumor volume. In the regression test, the only parameter that correlated with the operation time was tumor volume. In conclusion, we believe that tumor volumetric calculations may be useful in predicting difficult distal rectal carcinoma surgeries.

CNN-based plant disease recognition using colour space models

ABSTRACT The growing population increases the demand for agricultural production. Owing to some plant diseases, this production is continuously decreasing. Detecting and recognising these diseases at the initial stage is crucial for the sake of high yields from crops and plants. The conventional methods involve manual detection, which can be error-prone and laborious. Automatic recognition paves a way for the regular monitoring of huge areas of crops in less time. The deep learning-based approaches have entered in a big way in the detection of diseases in plants and crops because of their ability to overcome the limitation of regular manual monitoring. The use of colour space models helps demarcate the affected area ofimages of plants and crops prior to inputting them to the CNN model so as to improve the detection accuracy. For this, different colour space models that include RGB, HSL, HSV, LAB, LUV, XYZ and YUV are investigated. However, HSL-CNN joint model performs the best by achieving 98.97% accuracy, 98.61% precision, and 99.28% F1-score.

Three-Stage Deep Learning Framework for Video Surveillance

The escalating use of security cameras has resulted in a surge in images requiring analysis, a task hindered by the inefficiency and error-prone nature of manual monitoring. In response, this study delves into the domain of anomaly detection in CCTV security footage, addressing challenges previously encountered in analyzing videos with complex or dynamic backgrounds and long sequences. We introduce a three-stage deep learning architecture designed to detect abnormalities in security camera videos. The first stage employs a pre-trained convolutional neural network to extract features from individual video frames. Subsequently, these features are transformed into time series data in the second stage, utilizing a blend of bidirectional long short-term memory and multi-head attention to analyze short-term frame relationships. The final stage leverages relative positional embeddings and a custom Transformer encoder to interpret long-range frame relationships and identify anomalies. Tested on various open datasets, particularly those with complex backgrounds and extended sequences, our method demonstrates enhanced accuracy and efficiency in video analysis. This approach not only improves current security camera analysis but also shows potential for diverse application settings, signifying a significant advancement in the evolution of security camera monitoring and analysis technologies.

IoT Based Smart Agriculture / Gardening

Traditional agriculture rely heavily on manual monitoring and irrigation, leading to water waste, uneven crop growth, and increased labor costs. This paper proposes an Internet of Things (IoT)-based smart agriculture system for optimized irrigation. The system uses an ESP32 microcontroller to collect soil moisture data from sensors. User-defined thresholds trigger automated watering using a relay-controlled mechanism. The system offers on-site monitoring through an OLED display and remote monitoring capabilities via the Blynk platform. This paper details the system design, implementation, and testing, demonstrating its power for efficient water use and improved crop yield in agricultural and gardening applications.