Study On Application Of Optics Navigation In BRT (Bus Rapid Transit) System

Nowadays, science and technology advance by leaps and bounds. Here, in China, citilization is underway vigorously. However, the development of transportation cannot catch up with the high-speed development of the process of citilization. On the other hand, with more and more private cars, there is less and less room for transportation in our cities. The key to this problem is to vigorously develop public transportation. Bus Rapid Transit raised in some South American countries which have similar situations to China provide us with a good example in this aspect.In BRT system, vehicle moving speed is really fast, how to ensure that vehicles will be moving along the road that has been planned and how to navigate vehicles while vehicles in a high speed has become a very important issue in BRT.In this paper, we advanced the principle and algorithm using computer vision and pattern recognition in the computer vision navigate system of BRT, and give the specific application example. This dissertation is based on special road rights of Bus Rapid Transit and road tracking navigation line's particularity, at condition of the self-driving speed that's under 60k/h, carry through vehicle to road tracking navigation tracking navigation control. We also discussed how to navigate in BRT system under structural road environment in this paper. According to the character of the length of buses in BRT, we defined a new assembling method of vision sensor and the concept of structural road and the axiom model of structural, and proposed the relation of mutual converting road and image coordinate system. We placed three image collection units under the bus middle axis line, used the relation of mutual converting road and image coordinate system, fitting two curved line in the two different coordinate system, thus we can track and navigate the bus in BRT.In the unit of image pretreatment and the drawing of road tracking line, the system preprocess the image transmitted from the video camera sensors, eliminate the noise of image. This dissertation also introduced two drawing algorithms of road tracking line, based Hough transforming and curved line fitting. Traditionally, there are many kinds of road logo line drawing algorithms, according as the algorithms process the area of image, we can classify into partial and overall calculatingalgorithms. In the application of BRT, we adopted the partial algorithm used three image collection points, and in system combined the partial information, fitting the overall curved line of the road logo line.