研究目的
To propose a new image processing scheme for smart cars that can identify different types of tracks, filter out noise and interference, identify obstacles, and extract the available left and right edges of the track.
研究成果
The proposed new scheme can effectively extract and classify the image features of different road types and provide corresponding image processing algorithms. It is very effective in the fast conversion of image data, information extraction, noise cancellation, and image enhancement. Thus, the proposed scheme has a high practical value in applications, such as smart car contests and self-driving car industry.
研究不足
The computational intelligence of the embedded system limits the algorithm's complexity, requiring real-time and lightweight characteristics. Many existing bitmap processing algorithms cannot meet the demand.
1：Experimental Design and Method Selection:
The study involves designing a new image processing scheme for smart cars, focusing on track identification, noise filtering, and obstacle detection. The methodology includes image preprocessing, feature extraction, and classified recognition.
2：Sample Selection and Data Sources:
The study uses video images from the NXP smart car competition as the primary data source.
3：List of Experimental Equipment and Materials:
A digital camera is used to acquire gray-scale images, which are then converted into binary images based on the average brightness of the whole image using hardware.
4：Experimental Procedures and Operational Workflow:
The process includes bitmap preprocessing, search connectivity area, fixed-scale filtering, continuity search of the edge, feature extraction, and classified recognition.
5：Data Analysis Methods:
The study employs various algorithms for image processing, including erosion, dilation, open and close operations, cross-edge search algorithm, and fixed-scale filtering.
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
