A Survey on Machine Parts Recognition and Defect Detection using Image Processing Techniques R.Anusuya 1 , Dr .P .Arjun 2 , R.Haripriya 3 , Dr.C.Paulraj 4 1 Department of Computer Science and Engineering, University College of Engineering Villupuram 1 anurajendran1997@gmail.com Abstract - In this survey paper we have studied and organized the previous works related to machine parts processing using computer vision techniques. The survey mainly focuses on three aspects namely object recognition, defect detection and shape feature extraction. Object recognition can be used in different fields such as industries, robotics, medical diagnosis, video surveillance etc. Defect detection is used in quality control units toremove defective components to maintain the quality of products. Feature extraction begins with data acquisition and builds features they are informative and non-redundant, facilitating the subsequent learning and generalization steps, and in some cases leading to better understanding than human interpretations. Keywords – Defect Detection, Feature Extraction, Object Recognition, and Machine part Recognition, Computer Vision. I. INTRODUCTION Object recognition and defect detection in real world is easy for human beings. A person can recognize an object with little effort and having different viewpoints. Implementation on machines has been very difficult in Algorithm basis. A schematic diagram for an assembly line to recognize and defect detection is shown in Figure. 1. Fig.1 Machine Parts Defect Detection in Assembly line Object Recognition in terms of machine parts image processing is categorized into two phases, part recognition and defect detection. There are various challenges in object recognition, they have to be detected under noisy background, different lighting condition and contrast environment. There are many methods for defect detection they are methods based on statistical approach, model based methods, filtering based methods, spectral approaches and learning based approaches [17].Feature detection is a low-level image processing operation; it examines every pixel to see if there is a feature present at that pixel. Feature extraction can be categorized into two steps as Feature construction and feature selection. The rest of the paper has been organized in to four sections. Section 2 gives details about machine part recognition and detection. The section3 gives brief about defect detection methods. The feature extraction techniques are described in section4. The section 5summarizesthe paper and gives future work directions. II. Machine Part Recognition and Detection Machine Part recognition is a computer vision approach for identifying shapes and objects in images and it detects any part variation, distortion or noise found in the images. JASC: Journal of Applied Science and Computations Volume VI, Issue III, March/ 2019 ISSN NO: 1076-5131 Page No:1