Multiple-camera System for 3D Object Detection in Virtual Environment using Intelligent Approach

This work is the development of an adaptive multi-camera system for object detection which can be further adjusted for computational requirements. To get preliminary results, the whole project is simulated in a virtual environment for tracking the targeted object and to collect maximum information about the object’s features and characteristics. The virtual scene is developed using Unity software, and the images are captured by a virtual camera which is controlled by an external device (mouse). Further expected results of this project are to achieve a model which can be trained on the various classes of the 3D objects to classify objects and calculate the optimal distance between the object and the camera position, making it reliable to get the maximum information from the real-time environment which is developed by physical model by using camera and microcontroller raspberry pi4. The initial results obtained in this preliminary work clearly represent what has been achieved so far. Machine learning algorithms have been used to recognize what the object is that is in the camera’s gaze. Objects have been classified into different categories. Object detection has been implemented using Python and Opencv library.