|
| 基于神经网络的双目视觉传感器建模 |
| Binocular vision sensor modeling based on neural network |
| |
| DOI： |
| 中文关键词: 神经网络,视觉传感器,立体视觉,建模 |
| 英文关键词:neural network,vision sensor,stereoscopic vision,modeling |
| 基金项目: |
| 汪晓东 叶美盈 王霞 王辉 |
浙江师范大学数理与信息科学学院 浙江金华321004
(汪晓东，叶美盈)
，上饶师范学院物理系 江西上饶334001
(王霞)
，浙江师范大学数理与信息科学学院 浙江金华321004(王辉)
|
| 摘要点击次数: 2347 |
| 全文下载次数: 14 |
| 中文摘要: |
| 根据双目视觉传感器的工作原理 ,分析了影响测量精度的因素 ,表明双目视觉传感器的物体空间坐标与图像坐标之间存在复杂的非线性映射关系 ,其数学模型无法用解析式精确地加以描述。因此 ,提出一种基于神经网络的双目视觉传感器建模方法。数值实验表明 ,该方法简便易行 ,建立的模型有较高的精度 |
| 英文摘要: |
| The factors,which influence measuring precision of binocular vision sensor, are analyzed according to the sensor operating principle. The relation between the coordinates of the object and the coordinates of the image is the complex nonlinear mapping in the sensor. The mathematical model of the binocular vision sensor can't be accurately described by expressions.Therefore, this paper presents a new modeling method for binocular vision sensor based on neural network.The numerical experiment shows that the neural network is suitable for building up a nonlinear model of the binocular vision sensor. This method is simple, easy to realize. A higher measuring precision can be obtained. |
| HTML 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
