Real Time Road Edge Detection And Road Sign Recognition

Every year 1.3 million people worldwide are killed on roads and between 20 and 50 million are injured. A good solution to this problem would be to develop machines, which take into account the environment. That is why today, safe auto driving is becoming a popular topic in many fields, from small projects to large car factories. However, this topic also raises many questions and problems. There is a need to define the width of the edges of the road, recognize road signs, traffic lights, pedestrians, and other objects, which contribute the driving safely. There are many methods for solving these tasks.The experiments proposed in this paper based on the intellectual system of the vehicle. As experimental vehicle used smart car BJUT-IV, which is property of Beijing University of Technology. This paper discusses two of the most important tasks, which are detection edge of the road and traffic signs recognition.To start the work it needs to get the images from the camera and preprocess it. This part include changing the images to grey scale, filtering images and its transformation into binary image.The next step proposed some methods of transformation images to grey scale after what results of these methods compared. As the basis of the selected weighted average method. For image, filtering used to de-noise with application. To select a binarization method was done a large number of experiments. For this, we applied Otsu method.The next step after a successful image pre-processing, is detecting edges of the road. To do this, using the Hough method. To save time and resources, Hough method applied to the region of interest. Also using the implementation of Catmull-rom for the curves detection. Then the results of the two methods will applied for edges of the road detection that makes results more accurate.Quite time-consuming is the task of identifying the road signs that is need modernization. Task implemented by using SURF method. This method compares the abrupt changes of light in the images, noting the points as POIs, then the POI of two or more images comparing. This allows detecting objects with a good accuracy.During the calculations and experiments, there were some mistakes and disadvantages, which will discussed below.