Detection of people in an uncontrolled environment using the HaarCascade classifier

• Grant number: 20/06904-2
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Start date: July 01, 2020
• End date: June 30, 2021
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Celso Olivete Júnior
• Grantee: Leandro Henrique Lima e Silva
• Host Institution: Faculdade de Ciências e Tecnologia (FCT). Universidade Estadual Paulista (UNESP). Campus de Presidente Prudente. Presidente Prudente , SP, Brazil

Abstract

Detecting people in complex environments is a major challenge in the field of Computer Vision. This problem is accentuated when it comes to experiments in uncontrolled environments, containing several obstacles that substantially interfere in the result. Among the main factors that interfere in the process, and consequently in the result, we can mention the total or partial occlusion, the non-uniform illumination, the objects, the pose variation, the scale variation, among others. To mitigate these factors, pre-processing methods will be used to improve the quality of images, thus allowing a better performance of the detection algorithms. These steps (pre-processing and detection) will be applied in the project to be developed, which consists of the analysis of images obtained from surveillance cameras, with the objective of detecting the presence of students in classrooms and computer teaching laboratories, subsequently allowing the management of air conditioning units (automatically turning them off). The Python programming language will be used with the help of the library for the development of applications in the Computer Vision area – Open CV, which has modules to act in the pre-processing of images as well as classification (identification of people in the classroom/laboratory). computing) from classifying modules, in particular, in this project the classifier HaarCascade will be used. As a contribution of this project, it is hoped that from the analysis of images from surveillance cameras it will be possible to determine the detection of people in the environments already mentioned, allowing to provide data for a future project, in which computational techniques will be studied and used to allow the automatic management (on/off) of the air conditioning units and lamps, as well as that this management can be done remotely, aiming to reduce expenses and waste of electric energy. (AU)