International Journal Of Scientific Research And Education Volume 2 Issue 3 Pages 553-562 2014 ISSN (e): 2321-7545 Website: http://ijsae.in Colour And Shape Based Object Sorting Abhishek Kondhare, 1 Garima Singh 2, Neha Hiralkar 3, M.S.Vanjale 4 AISSM s Institute Of Information Technology. Dept. Of Electronics Engineering, Pune, India. Email-k.abhisheknow@gmail.com Abstract The project is a smart approach for a real time inspection and selection of objects in continuous flow. Image processing in today s world grabs massive attentions as it leads to possibilities of broaden application in many fields of high technology. The real challenge is how to improve existing sorting system in the modular processing system which consists of four integrated stations of identification, processing, selection and sorting with a new image processing feature. Existing sorting method uses a set of inductive, capacitive and optical sensors which do differentiate object colour. This project presents a mechatronics colour sorting system solution with the application of image processing. Image processing procedure senses the objects in an image captured in real-time by a webcam and is classified using a decisional algorithm and selected in real time..this information is processed by image processing for pick-and-place mechanism. This project uses an automated material handling system which is widely used in industries. 1.INTRODUCTION This chapter will briefly discuss on the project background and its scope. The primary objective of our project is to make a system which sorts predefined objects using image processing on the basis of colour and shape. The system comprises of essential units such as conveyer unit, camera unit, robotic arm and IR sensor unit. Robotics field not only deals with electronic circuits but with mechanical basics also. Hence we have chosen robotics for pick and place mechanism. Determining real time and highly accurate characteristics of small objects in a fast flowing stream would open new directions for industrial sorting processes. This system relates to an apparatus and method to classify and sort objects, using electronic systems and advanced sensors operating on the basis of a physical and geometric characterization of each Abhishek Kondhare et al IJSRE Volume 2 Issue 3 March 2014 Page 553
element. Recent advances in electronics and printed circuit board technology open new perspectives for industrial application in this field. Sorting system is one of the useful systems in Industries. It can be used to differentiate item based on the color and shape of the item itself. This system is implemented to improve the process of the industry. Traditionally, the object sorting process was done by the operator manually. However, this method has some disadvantages such as increase in the cost of the product, slow, and inaccuracy due to the human mistake. Color and shape based Sorting System with Robot Arm is a solution to this problem. This system will be operated using Programmable Interface Controller ( PIC). For color & shape detecting, it will use the webcam which is interfaced with microcontroller using USB. The most significant part of this project is to have a robotic arm. This robot arm s function is to pick and place and its gripper can move in a circular path. Since this system is mainly controlled by the PIC microcontroller, the result of the sorting process will be more reliable and faster. 2.SYSTEM DESIGN AND METHODOLOGY Fig 1: System Block Diagram The proposed system works in following three steps: 2.1. Image Acquisition 2.2. Image Processing 2.3. Sorting Mechanism Abhishek Kondhare et al IJSRE Volume 2 Issue 3 March 2014 Page 554
2.1 IMAGE ACQUISITION To start with when the object on the conveyer is detected by the IR sensors, image is captured by the camera and is sent to the MATLAB workspace. The input image from the camera cannot be given for processing directly. Pre-processing is done on the image such as thresholding. Then only object image I converted in binary format. Then this binary image has to be smoothened so that paper and salt noise can be removed. This final threshold image of object is now ready for processing[1]. 2.2 IMAGE PROCESSING The objects are sorted on the basis of colour and shape. For shape recognition captured image is converted to gray from RGB. Then thresholding is done followed by inverting the image. Boundaries Concentrate are found and lastly shape is identified using shape properties [2]. To identify the color, firstly the image is coverted into gray format and then thresholding is done. After thresholding colour components are extracted and the image is converted into black and white format which is called as binary format. Remove all those pixels less than 300px. Label all the connected components in the image. Find region properties & bounding box and the color is identified [3]. 2.3 SORTING MECHANISM The sorting mechanism consist of a robotic arm and a conveyer assembly. After calculating the size and identifying the color, command will be sent to direct the motor of a robotic arm through com port of the computer. Conveyor assembly is in OFF state for this period. According to the size and color the robotic arm places the objects. 3. HARDWARE DESIGN TECHNIQUE 3.1 Power supply design 3.2 Interfacing of camera 3.3 Interfacing of placing system Abhishek Kondhare et al IJSRE Volume 2 Issue 3 March 2014 Page 555
3.1 POWER SUPPLY DESIGN Fig.2 Power Supply The bridge rectifier and capacitor i/p filter produce an unregulated DC voltage which is applied at the I/P of 7805.As the minimum dropout voltage is 2v for IC 7805, the voltage applied at the input terminal should be at least 7 volts. C1 (1000 µf / 65v)is the filter capacitor and C2 and C3 (100n f) is to be connected across the regulator to improve the transient response of the regulator. Assuming the drop out voltage to be 2 volts, the minimum DV voltage across the capacitor C1 should be equal to 7volts (at least). 3.2 INTERFACING OF CAMERA 12 megapixel night-vision camera which is having USB interfacing facility is used to capture the image. The camera is directly interfaced to USB port. Interfacing of placing system. 3.3 INTERFACING OF SORTING SYSTEM Fig.3 DC motor interface Abhishek Kondhare et al IJSRE Volume 2 Issue 3 March 2014 Page 556
Here in our project we are using a 12v DC motor which is Bipolar that means the DC motor can rotate in both the sides. For this we are using a DC motor driver IC L293D. This driver IC can drive 2 DC motors. The DC motor will be connected at OUT1 and OUT2 of L293D respectively. Features of Driver IC L239Di. 600mA output current capability per channel. ii. Over Temperature Protection. iii. High noise Immunity. 4. SOFTWARE DESIGN FLOWCHART Fig. 4 Colour Sorting flowchart Abhishek Kondhare et al IJSRE Volume 2 Issue 3 March 2014 Page 557
Fig.5 Shape Sorting flowchart Abhishek Kondhare et al IJSRE Volume 2 Issue 3 March 2014 Page 558
5. ENTIRE SYSTEM OPERATION Object sorting is done using two ways: Color based sorting and shape based sorting. 5.1 COLOUR BASED SORTING After image acquisition thresholding is done. Then red, green and blue colour pixels are calculated. The image is converted into binary. Remove all those pixels less than 300px. Label all the connected components in the image. Find region properties & bounding box and Find red or green or blue object. 5.2 SHAPE BASED SORTING Acquire the image. Convert that image from rgb to gray. Threshold the image. Convert it to the Binary Image and then invert it. Find the boundaries Concentrate. Determine Shapes properties. Classify Shapes according to properties 6. APPLICATIONS 1) In small scale and large scale industries to sort the products based on the various parameters. 2) Can be used in departmental store. 3) In malls and small shops. 4) In various industries to sort the bottles of various sizes such as medicine and wine industry. 5) In food industries to identify the rotten or damaged fruits. 6) Artificial robotic intelligence. 7) In bio medical field, color and shape analyzing algorithm can be used for recognition of cancer cells. 7. RESULT AND DESCUSSION The output module consist of (a)conveyer Model (b)robotic Arm Abhishek Kondhare et al IJSRE Volume 2 Issue 3 March 2014 Page 559
Fig.6 Conveyor and gripper assembly Fig.7 Gripper Abhishek Kondhare et al IJSRE Volume 2 Issue 3 March 2014 Page 560
Fig.8 Objects of different shapes and colours Fig.9 Fig.10 Abhishek Kondhare et al IJSRE Volume 2 Issue 3 March 2014 Page 561
8. Conclusion Thus we have developed colour and shape based object sorting system using image processing. This is a user friendly model which uses robotic arm mechanism for sorting and a webcam for taking images of the various objects. It also includes the miniature version of conveyer module. The result of the system abducts due to environmental conditions but it can be enhanced by improving camera quality. ACKNOWLEDGMENT We would like to thank our project guide Prof. M.S.Vanjale, H.O.D. Prof. Shedge for helping us in writing this paper. REFERENCES [1] Rafael C. Gonzalez and Richard E. Woods, Digital Image Processing, Third Edition, Vol.3, Pearson Publication. [2] www.mathworks.com [3] Alessandro Golfarelli, Rossano Codeluppi and Marco Tartagni, A Self-Learning Multi-Sensing Selection Process: Measuring Objects One by One,ARCES LYRAS LAB University of Bologna, Campus of Forlì, 1-4244-1262-5/07/$25.00 2007 IEEE, IEEE SENSORS 2007 Conference. [4] The 8051 microcontroller and embedded systems (second edition) by Mohammad Mazidi, Janice G. Mazidi & Rollin D. Mckinlay. [5] Power Electronics by M.D. Khanchandani, K.B.Singh(1998) Abhishek Kondhare et al IJSRE Volume 2 Issue 3 March 2014 Page 562