Image Segmentation - Region Growing Algorithm

1 - Introduction and problem definition

1.1 - Introduction

Image segmentation is an important process in Computer Vision that is used for several operations as edge detection, classification, 3D reconstruction, etc.. The main goal of image segmentation is to cluster pixels into regions. This clustering image pixels into image regions in turns convert the image into a representation that is more meaningful and easier to analyse. More precisely, image segmentation is the process of assigning a label to every pixel in an image such that pixels with the same label share certain characteristics. Image segmentation has broad range of applicability in different fields of science ans engineering.

In this lab work, we implement the region growing algorithm which is one of the basic process of partitioning a digital image and then analyse the design and implementation of it. Finally, we compare the region growing algorithm with other image segmentation algorithms. We describe also about the organization and development phase of the lab work.

1.2 - Problem definition

Our lab work problem asks for performing image segmentation over different image representation and check the result. We implement our image segmentation algorithm over gray level images and RGB color space images to cluster into different image regions. And then we compare our clustering result with Fuzzy C-Means (FCM) clustering algorithm.

Implementing Object Detection Based on Color in Webot Simulator for E-puck

This project was implemented by Richa AGARWAL, Taner GUNGOR and Pramita WINATA.

Abstract-Object detection and recognition is a challenging task in computer vision systems. So it was decided to work with E-puck for the same. But using a real e-puck connected with the system through bluetooth it is diffucult to transfer images captured by the robot's camera. So, it was decided to use Webot simulator for E-puck robot to develope and test the algorithm to detect objects using the color of an object. Where robot scans for the object if detects the goal, it moves in the direction of goal avoiding obstacles, else moves randomly in the arena looking for the goal (red object). The most relevant aspects of the simulator and implementation are explained.
Keywords-Webot simulator, e-puck, path planning

INTRODUCTION
We are implementing a simple object detection algorithm in Webot simulator for E-puck using C controller. The algorithm is designed to detect red objects using E-puck's camera. It is easier to control and grab images from E-puck robot using Webot simulator and controler.

1 - WEBOTS SIMULATOR
Webots is a development environment used to model, program and simulate mobile robots. With Webots the user can design complex robotic setups, with one or several, similar or different robots, in a shared environment. The properties of each object, such as shape, color, texture, mass, friction, etc., are chosen by the user. A large choice of simulated sensors and actuators is available to equip each robot. The robot controllers can be programmed with the built-in IDE or with third party development environments. The robot behavior can be tested in physically realistic worlds. The controller programs can optionally be transferred to commercially available real robots. Webots is used by over many universities and research centers worldwide. The development time you save is enormous.

Figure-1: Webots development stages

Webots allows you to perform 4 basic stages in the development of a robotic project Model, Program, Simulate and transfer as depicted on the Fig. 1.

Understanding k-Nearest Neighbour

Goal

In this chapter, we will understand the concepts of k-Nearest Neighbour (kNN) algorithm.

Theory

kNN is one of the simplest of classification algorithms available for supervised learning. The idea is to search for closest match of the test data in feature space. We will look into it with below image.

In the image, there are two families, Blue Squares and Red Triangles. We call each family as Class. Their houses are shown in their town map which we call feature space. (You can consider a feature space as a space where all datas are projected. For example, consider a 2D coordinate space. Each data has two features, x and y coordinates. You can represent this data in your 2D coordinate space, right? Now imagine if there are three features, you need 3D space. Now consider N features, where you need N-dimensional space, right? This N-dimensional space is its feature space. In our image, you can consider it as a 2D case with two features).

Now a new member comes into the town and creates a new home, which is shown as green circle. He should be added to one of these Blue/Red families. We call that process, Classification. What we do? Since we are dealing with kNN, let us apply this algorithm.