Ecosystem evaluation (1989-2012) of Ramsar wetland Deepor Beel using satellite-derived indices.

The unprecedented urban growth especially in developing countries has laid immense pressure on wetlands, finally threatening their existence altogether. A long-term monitoring of wetland ecosystems is the basis of planning conservation measures for a sustainable development. Deepor Beel, a Ramsar wetland and major storm water basin of the River Brahmaputra in the northeastern region of India, needs particular attention due to its constant degradation over the past decades. A rule-based classification algorithm was developed using Landsat (2011)-derived indices, namely Normalised Difference Water Index (NDWI), Modified Normalised Difference Water Index (MNDWI), Normalised Difference Pond Index (NDPI), Normalised Difference Vegetation Index (NDVI) and field data as ancillary information. Field data, ALOS AVNIR and Google Earth images were used for accuracy assessment. A fuzzy accuracy assessment of the classified data sets showed an overall accuracy of 82 % for MAX criteria and 90 % for RIGHT criteria. The rules were used to classify major wetland cover types during low water season (January) in 1989, 2001 and 2012. The statistical analysis of the classified wetland showed heavy manifestation in aquatic vegetation and other features indicating severe eutrophication over the past 23 years. This degradation was closely related to major contributing anthropogenic factors, such as a railway line construction, growing croplands, waste disposal and illegal human settlements in the wetland catchment. In addition, the landscape development index (LDI) indicated a rapid increase in the impact of the surrounding land use on the wetland from 1989 to 2012. The techniques and results from this study may prove useful for top-down landscape analyses of this and other freshwater wetlands.

Real-Time Remote Patient Monitoring and Alarming System for Noncommunicable Lifestyle Diseases.

Telemedicine and remote patient monitoring (RPM) systems have been gaining interest and received adaptation in healthcare sectors since the COVID-19 pandemic due to their efficiency and capability to deliver timely healthcare services while containing COVID-19 transmission. These systems were developed using the latest technology in wireless sensors, medical devices, cloud computing, mobile computing, telecommunications, and machine learning technologies. In this article, a real-time remote patient monitoring system is proposed with an accessible, compact, accurate, and low-cost design. The implemented system is designed to an end-to-end communication interface between medical practitioners and patients. The objective of this study is to provide remote healthcare services to patients who need ongoing care or those who have been discharged from the hospital without affecting their daily routines. The developed monitoring system was then evaluated on 1177 records from MIMIC-III clinical dataset (aged between 19 and 99 years). The performance analysis of the proposed system achieved 88.7% accuracy in generating alerts with logistic regression classification algorithm. This result reflects positively on the quality and robustness of the proposed study. Since the processing time of the proposed system is less than 2 minutes, it can be stated that the system has a high computational speed and is convenient to use in real-time monitoring. Furthermore, the proposed system will fulfil to cover the lower doctor-to-patient ratio by monitoring patients from remote locations and aged people who reside in their residences.