Human Activity Identification And Recognition In Aerial Images Using CNN Deep Learning Models

Abstract

The Human Action Recognition (HAR) from images, video clips is motivated due to abundant availability of videos and images. There is a need of diverse applications for automatic observation of sick people, security of senior citizens and kids through interfaces between man and machine. Automatic labeling of activities of people in images captured through the cameras of closed circuit television will be the motive of such applications. In recent years there is huge demand of image analysis and in particular -tagging of suspicious actions in surveillance system. The main objective of this work is to develop a model that can identify and classify the predefined activities of an individual. In addition to its technical significance, this research also recognizes the broader relevance of sustainable management practices and their transformative role in environmental sciences. Integrating energy-efficient computational models and responsible data handling mechanisms can contribute to reduced resource consumption and improved ecological outcomes. Moreover, by exploring how management strategies, innovations, and policies influence the deployment of HAR systems, it becomes possible to align technological progress with environmental sustainability goals across sectors and societies. This dual perspective not only highlights the utility of HAR in improving human welfare and security but also emphasizes its potential in supporting global sustainability frameworks.