研究目的
To design a system to detect and provide driver safety by alerting the driver of upcoming traffic signs using visual features like shape and color detection implemented by using OpenCV and Emgu with Visual Studio 2013.
研究成果
The proposed method of traffic sign detection using contour analysis approach with BGR to HSV conversion model and morphological filter for noise filtering is robust and can help in prevention of accidents by alerting the driver of road signs via audio message. The system shows 80 percent accuracy even with 90 degrees of rotation in sign boards.
研究不足
Contour analysis fails at identifying boards with separate contours as with the case with narrow bridge. The results are 80 percent appropriate even with 90 degrees of rotation in sign boards.
1：Experimental Design and Method Selection:
The proposed technique uses contour based analysis to detect traffic signs implemented using OpenCV and Emgu with Visual Studio 2013. Real-time objects are captured by the camera, converted into HSV from BGR color scale, noise filtered, and then used for extracting region-of-interest which is further classified into being road sign or not.
2：Real-time objects are captured by the camera, converted into HSV from BGR color scale, noise filtered, and then used for extracting region-of-interest which is further classified into being road sign or not.
Sample Selection and Data Sources:
2. Sample Selection and Data Sources: Real-time images captured by the camera.
3：List of Experimental Equipment and Materials:
Camera, OpenCV, Emgu, Visual Studio
4：Experimental Procedures and Operational Workflow:
20 Images undergo BGR to HSV conversion, noise filtering (Erode, Dilate, Smoothening, Thresholding), contour analysis, and comparison with a trained database to identify traffic signs.
5：Data Analysis Methods:
Contour analysis is used to explicate, store, scrutinize and find the object which are represented as exterior outline.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
