REVIEW
For map data is very important, there is no data, there is no map service. Users in the use of map services, data less likely to think like an iceberg, visible to the user only the most direct and most dominant part of product features, and support the foundation of the dominant part of the need, often much larger.
The map data was first collected from the professional to the acquisition tool is the cars, bicycles, aircraft and satellite imagery, the last two years has been the use of intelligent hardware crowdsourcing acquisition. After collection, the data is updated to enhance the speed and accuracy are unlimited. Because the ground is changing so fast, users increasingly rely on mapping applications. Therefore, the accuracy of the data update rate and data is care about the user experience of the company's top priority of the map. The first step in updating the data, that is, traffic sign detection.
This article will focus on the specific application of machine learning techniques in high German map data production and design of these technical solutions have been validated, and achieved good results, and provides the basis for the rapid updating of technology to ensure high moral map data.
"Traffic sign detection" is defined
Traffic sign detection, especially on an ordinary street view image detected by various types of automated means