CNC & Robotics Presentation on Image Processing PRESENTATION ON IMAGE PROCESSING 06/11/2020 Praveen Ram K Pravesh A R 17M912 17M913 Prethev R Sriram B 17M914 17M915 Visahan R 17M916 Visweshwara K Elumalai P 17M917 17M918 15M701 BE MECH (SW) - 9 SEM 1 Department of Mechanical 1 Engineering - PSG CT CNC & Robotics Presentation on Image Processing CONTENTS • • • • • • • Machine Vision Key components Functions What can Machine Vision do for Robotics? Vision Sensors Image Processing Object identification. 06/11/2020 15M701 BE MECH (SW) - 9 SEM 2 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing MACHINE VISION ● Vision for robots requires the ability to identify and accurately determine the positions of all relevant three dimensional objects within the robot work place. ● Machine vision may be defined as the process of extracting, characterizing, and interpreting information from images of a three dimensional world 06/11/2020 15M701 BE MECH (SW) - 9 SEM 3 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing KEY COMPONENTS 06/11/2020 15M701 BE MECH (SW) - 9 SEM 4 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing FUNCTIONS 06/11/2020 15M701 BE MECH (SW) - 9 SEM 5 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing OVERVIEW 06/11/2020 15M701 BE MECH (SW) - 9 SEM 6 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing PURPOSE OF A MACHINE VISION SYSTEM • • The main purpose of the machine vision system is to provide input to the system and obtain output from the system. The arrived output is analysed and converted into standard format. 06/11/2020 15M701 BE MECH (SW) - 9 SEM 7 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing MACHINE VISION-FUNDAMENTAL TASKS ● Image transformation ● Image analysis ● Image understanding Fig.2 Block diagram 06/11/2020 15M701 BE MECH (SW) - 9 SEM 8 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing GENERAL PURPOSE MACHINE VISION ● System should capture the relevant data and with the motion of the object it should be able to update the information. ● Four steps to General Purpose Robot Vision ● Object verification and tracking ● Fast extraction of stable image features ● Object model acquisition ● Efficient indexing of the model database 06/11/2020 15M701 BE MECH (SW) - 9 SEM 9 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing WHAT CAN MACHINE VISION DO FOR ROBOTICS? ● Accurate Robot-Object Positioning ● Keeping Relative Position under Movement ● Visualization / Teaching / Tele robotics ● Performing measurements ● Object Recognition ● Registration Fig.4 robot arm 06/11/2020 15M701 BE MECH (SW) - 9 SEM 10 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing VISION SENSORS ● ● Single Perspective Camera Multiple Perspective Cameras (e.g. Stereo Camera Pair) ● Laser Scanner ● Omnidirectional Camera ● Structured Light Sensor Fig.5 Vision sensors 06/11/2020 15M701 BE MECH (SW) - 9 SEM 11 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing SINGLE PERSPECTIVE CAMERA ● Single perspective camera in robotics are used in places where only a single camera is used. ● It has a fixed field of vision and the main application of the camera would be object identification. ● The accuracy of the details is low in these kind of camera setup in robots. 06/11/2020 15M701 BE MECH (SW) - 9 SEM 12 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing MULTI PERSPECTIVE CAMERA ● Multi perspective camera systems are deployed in places where the accuracy of the details obtained from the visual is to be high. ● In several cases more than one camera is employed for achieving a wider field of vision. ● Addition of cameras improves the depth and location accuracy of the object in the field of view as different perspective angles are obtained. 06/11/2020 15M701 BE MECH (SW) - 9 SEM 13 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing LASER SCANNER ● Range measurements are based on the TOF ● ● ● principle, in which a pulse of light is emitted by the laser scanner, and the time taken for it to be returned is recorded. Using the known constant speed of light, the range to the target can be calculated. A rotating polygon with reflective surface is used for creating 3D images. Horizontal frame scans are then obtained by rotating the optical head. 06/11/2020 15M701 BE MECH (SW) - 9 SEM 14 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing LASER SCANNER 06/11/2020 15M701 BE MECH (SW) - 9 SEM 14 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing OMNIDIRECTIONAL CAMERA ● Omnidirectional cameras are frequently used for visual odometry .Due to its ability to capture a 360-degree view, better results can be obtained for optical flow and feature selection and matching. 06/11/2020 15M701 BE MECH (SW) - 9 SEM 16 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing OMNIDIRECTIONAL CAMERA ● In robotics and computer vision, visual odometry is the ● process of determining the position and orientation of a robot by analyzing the associated camera images. It has been used in a wide variety of robotic applications, such as on the Mars Exploration Rovers. 06/11/2020 15M701 BE MECH (SW) - 9 SEM 17 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing IMAGE PROCESSING ● ● ● ● Image data reduction Segmentation Feature extraction Object recognition 06/11/2020 15M701 BE MECH (SW) - 9 SEM 18 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing Image data reduction ● This method is used to reduce the volume of data by two methods 1.Digital conversion 2.Windowing Digital conversion: ● This method of volume reduction is done by reducing the no of grey levels used by machine vision system. Windowing: ● This method uses only a portion of image stored in frame buffer for image processing and analysis. Contrast ratios are reduced to limited pixels to save space. 06/11/2020 15M701 BE MECH (SW) - 9 SEM 19 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing IMAGE SEGMENTATION ● Division of image into multiple segments. In this process a label is assigned to every pixel in an image such that the pixel with same label share certain characteristics. Techniques of segmentation include 1.Thresholding 2.Region growing 3.Edge detection 4.Feature extraction 06/11/2020 15M701 BE MECH (SW) - 9 SEM 20 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing IMAGE SEGMENTATION ● ● Thresholding is a binary conversion technique in which pixels are converted into binary level such as black and white. Region segmentation is done by grouping the pixels into regions called grid elements based on attributes. 06/11/2020 15M701 BE MECH (SW) - 9 SEM 21 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing IMAGE SEGMENTATION ● Edge detection considers the change in intensity that occurs in the pixels as the ● boundary or edges of a part. It separates the part from the background. Feature extraction is single parameter where the features of parts such as area,diameter,perimeter,ridge detection etc. are identified and obtained from the image. 06/11/2020 15M701 BE MECH (SW) - 9 SEM Department of Mechanical 22 Engineering - PSG CT CNC & Robotics Presentation on Image Processing OBJECT RECOGNITION ● The task of finding and identifying objects in an image or video sequence. Pattern recognition: ● Pattern recognition focuses on recognition of patterns and irregularities in data. ● It is done by two methods 1.Template matching 2.Weighing feature. 06/11/2020 15M701 BE MECH (SW) - 9 SEM 23 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing TEMPLATE MATCHING ● It is used to for finding small parts of an image which match a template image ● It can be used in manufacturing as a part of quality control a way to navigate a mobile robot, or as a way to detect edges in images ● The features of the object that can be used in template pattern are area , diameter ,aspect ratio. 06/11/2020 15M701 BE MECH (SW) - 9 SEM 24 Department of Mechanical Engineering - PSG CT CNC & Robotics Presentation on Image Processing FEATURE WEIGHTING ● Several features like area, length and perimeter are combined into single measures by assigning a weight to each feature according to the relative importance in identifying the object ● The score of the object in the image is compared with the score of the computer memory to achieve proper identification 06/11/2020 15M701 BE MECH (SW) - 9 SEM 25 Department of Mechanical Engineering - PSG CT CNC & Robotics 06/11/2020 Presentation on Image Processing 15M701 BE MECH (SW) - 9 SEM 26 Department of Mechanical Engineering - PSG CT