1. Home

# Object Detection Using Asus Xtion Pro Live (3d sensor)

Detecting objects in 3D space is often a challenging task. The advent of 3d scanners (RGB-D cameras) and point cloud libraries have enabled robotic researchers to make an autonomous robot perform specific tasks with known objects. Though the field of Artificial Intelligence witnessed 5 decades of augmented research in computer vision, there remain not precise or generic solutions to detect or locate objects when it comes to the real world. The problem of object detection cannot be generalized and it is highly environmentally sensitive and domain specific. Here I present an implementable approach to detect and locate known, colored, cubic objects by the Baxter research robot with the help of the Asus Xtion Pro Live (3d scanner) sensor in the blocks-world domain. This work is limited to depicting the practical tractability of the cognitive architecture (MIDCA) in the blocks-world domain with Baxter robot and to further enhancing the capabilities of the robot to perform tasks in various other practical domains that are defined in MIDCA.

Introduction to the Method used for object detection