Autonomous Real-Time Navigation Based on Dynamic Line and Object Detection

Abstract

An Android application has been developed that captures images, detects a line and computes the lateral deviation as well as angular deviation from the path. It also takes into consideration the intensity variation of the surroundings which has been explained in the following paper. We have used the Sobel edge detection algorithm and Hough transform for the angular inference. Colour thresholding and matrix manipulation are done for calculating the lateral deviation. Finally, the counter-active actions and inferences are sent via BLE (low energy Bluetooth). An application was developed for directing the autonomous quadruped robot. The rare-view camera of the mobile was used for taking the input. This Android application can have multiple implementations apart from robotics with slight manipulations, such as the automobile industry, military, industrial automation and blind assistance, and these kinds of researches will boost concept of smart manufacturing.