International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056 Volume: 07 Issue: 05 | May 2020 www.irjet.net p-ISSN: 2395-0072 © 2020, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 2093 Vision Based Road Hump and Speed Breaker Detection Prof. Varunakshi Bhojane 1 , Romali Surve 2 , Krunal Rane 3 1 Professor, Dept of Computer Engineering, Pillai College of Engineering, New Panvel, Maharashtra, India 2,3 B.E. Computer Engineering Student, Pillai College of Engineering, New Panvel, Maharashtra, India ---------------------------------------------------------------------***--------------------------------------------------------------------- Abstract - Speed Breaker detection and tracking are widely applied to intelligent image and video surveillance. The goal is to detect Speed Breaker in a scene, localize them in the image and, if possible, infer their exact articulations. This research addresses the deep learning–based techniques. By using the You Only Look Once version 2 (YOLOv2) detector and deep convolutional neural network (CNN) we detect the objects obstructing while driving the vehicle. Poor road conditions like cracks, potholes, open sewer caps, and extreme road conditions can cause inconvenience to passengers, damage to vehicles, and accidents. Detecting those obstacles has become relevant due to the rise of the autonomous vehicle. Although previous studies used various sensors and applied different image processing techniques, performance is still significantly lacking, especially when compared to the tremendous leaps in performance with computer vision and deep learning. Hence, the object detection technique (YOLO) will be used. We do not make a selection or technique best method and optimal because the best technique depends on the needs, concerns and existing environment. The goal of object detection is to detect all instances of objects from a known class, such as people or faces in an image. Typically, only a small number of instances of the object are present in the image, but there is a very large number of possible locations and scales at which they can occur. Key Words: Speed Breaker detection, Computer vision, Machine Learning, Deep Convolutional, Neural Network 1. INTRODUCTION The quality of road is crucial to people who drive. In some areas, drivers need to be cautious because damaged speed breaker and road humps have been proven to cause catastrophes, especially during the rainy season. Detecting humps would allow vehicles to issue warnings so drivers can slow down and avoid them (or the vehicle itself can adjust settings to avoid them), minimize the impact, and make the ride smooth. Speed breaker detection is about sensing the road ahead of an autonomous vehicle. Nonetheless, studies and research on road-surface damage are still few. Several of them use traditional methods, with sensors and expensive equipment to label images in a classification task, but not to detect dam- age coordinates. Recently, object detection using end-to-end deep learning has been reported to outperform traditional methods. Costly sensors, battery life, computation power, and the complexity of data integration have been reduced by simply relying on imagery input to detect objects. In this study, we train and evaluate object detection with You Only Look Once version 2 (YOLOv2) that has a state-of-the-art convolutional neural network (CNN) at its core.[1] Literature survey describes various techniques used in the work. Identifying the current literature in related domain problem and identifying the techniques that have been developed and present the various advantages and limitation of these methods used extensively in literature. A) Detection of Potholes Using a Deep Convolutional Neural Network - Although previous studies used various sensors and applied different image processing techniques, performance is still significantly lacking, especially when compared to the tremendous leaps in performance with computer vision and deep learning. This research addresses this issue with deep learning–based techniques. We applied the You Only Look Once version 2 (YOLOv2) detector and pro- pose a deep convolutional neural network (CNN) based on YOLOv2 with a different architecture. Despite a limited amount of learning data and the challenging nature of pothole images, our proposed architecture is able to obtain a significant increase in performance over YOLOv2 (from 60.14 to 82.43 average precision).[1] B) Real Time Detection of Speed Hump/Bump and Distance Estimation with Deep Learning using GPU and ZED Stereo Camera - Most of the humps in India are n o t being constructed and maintained according to the public safety guidelines of Indian Road Congress (IRC) i.e., IRC099, which is resulting in damage to the vehicles, severe discomfort to the driver and even causing loss of direction control which is leading to fatalities. Very few methods were discussed in literature for un-marked speed hump/bump detection. We propose a method that detects and informs the driver about the upcoming un-marked and marked speed hump/bump in real time using deep learning techniques and gives the distance the vehicle is away from it using stereo- vision approaches.[2]