Integrated Framework for Identification of Suspicious Activity for Remote Sensing Applications
R. Gupta, Sachin Kumar, Pooja R. Khanna, Pragya Pragya · 2023
With exponential growth in number of connected devices, interconnected networks have grown multi fold, this has led to ease with digital transactions can happen among different domains of industries, be it finance, surveillance, transport, manufacturing, food, agriculture, healthcare or any other. Development aforesaid has potentially put forward an open challenge in terms of confidentiality, security, and integrity resulting in attacks be it physical or in online domain, more than ever owing to number of details and intricacies visible with ease. Potential threats to security have taken a new form aided with innovations in technology, which has boosted the need to upgrade and revamp counter systems to ensure optimal security and has also opened several potential research domains. Most of the existing surveillance systems are employed for investigations only after an incident has taken place and generally, inputs from cameras are monitored by human operators or stored for later use and are inspected if the need arises. With multiple inputs the task becomes too cumbersome, number of solutions have been proposed to address the issue, however a wide gap still exists. Work presents a framework for detecting suspicious vehicle movement, presence of unauthorized bunker, and suspicious human activity in military areas, employing data from remote sensing sites and ArcGIS an open platform for acquiring remote data. A data set of 1326 images were employed for the proposed work with 380 images for vehicle movement, 514 images for bunker detection, and 423 images for human activity. Input to proposed model is a combination of remote sensing and ArcGIS data. Model proposed classifies the data into groups as suspicious that needs attention and non-suspicious employing AI algorithms. Model achieves an accuracy of 91.65%, precision of 91.76%, recall of 91.65 and f1 score 91.64.