|
| 基于STM32和K210视觉识别的智能车设计与实现 |
| Design and Implementation of an Intelligent Vehicle Based on STM32 and K210 Vision Recognition |
| 投稿时间:2025-09-16 |
| DOI:10.16018/j.cnki.cn32-1650/n.202504008 |
| 中文关键词: STM32 视觉识别 智能车 |
| 英文关键词: STM32 visual recognition intelligent vehicle |
| 基金项目: |
|
| 摘要点击次数: 46 |
| 全文下载次数: 22 |
| 中文摘要: |
| 自动驾驶技术是智能汽车发展的核心方向,其中基于视觉感知的环境理解方案因成本低、信息丰富而备受关注。针对结构化场景下的路标识别与自主控制问题,设计并实现了一种以STM32微控制器为控制核心,K210为视觉处理单元的智能车系统。系统通过摄像头采集道路图像,由K210模块基于YOLOv2目标检测算法实时识别转向、限速与红绿灯等的关键路标,通过串口传输至STM32,进而控制电机完成循迹、转向与调速等动作。实验结果表明,系统在K210平台上实现约14帧/s的稳定运行,平均检测精度达91.5%(mAP@0.5),在模拟路况下实现了基本的自动驾驶功能。 |
| 英文摘要: |
| Autonomous driving technology represents a core direction in the development of intelligent vehicles. Among various approaches, vision-based environmental perception solutions have attracted significant attention due to their low cost and rich informational output. To address the challenges of road sign recognition and autonomous control in structured environments, this paper designs and implements an intelligent vehicle system featuring an STM32 microcontroller as the control core and a K210 visual processing unit for vision tasks.The system captures road images via a camera, and the K210 module employs the YOLOv2 object detection algorithm to recognize key road signs—such as turn indicators, speed limits, and traffic lights—in real time. The recognition results are transmitted via serial communication to the STM32, which then controls the motors to perform actions including path following, steering, and speed adjustment. Experimental results show that the system operates stably on the K210 platform at approximately 14 f/s, achieving a mean average precision (mAP@0.5) of 91.5%. It successfully demonstrates basic autonomous driving capabilities under simulated road conditions. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|