weixin_38530678的博客

Only one localization algorithm indoor environment has certain error, and the change of localization environment will cause instability of positioning system. The fusion of the position fingerprint matching algorithm and the polynomial distribution model can reduce the influence of the low positioning accuracy caused by the shortcomings of the polynomial distribution model and the position fingerprint matching model. In this paper, the position fingerprint matching algorithm and the polynomial distribution algorithm are respectively used to locate in different environments, and the same parameter is used to quantify the positioning results of the two different algorithms on the same environment. According to the selection coefficient, the optimal algorithm is selected for indoor positioning. In the online positioning stage, an algorithm that can be selected according to the selection coefficient to adapt to the environment can be used to locate. This adaptive algorithm can solve the respective defects of the fingerprint matching algorithm and the polynomial distribution model, and improve the indoor positioning accuracy.

1、在对相机标定的研究中， 以张正友相机标定法为理论基础， 利用采集到的标定板图像，通过单应性矩阵的计算， 结合 OpenCV 完成相机标定， 得到相机内参和畸变参数； 2、对当前的基于机器视觉的车轮定位技术进行研究的前提下， 提出了基于单目视觉的车轮定位技术方案， 将方案分为图像采集和数据处理两个部分， 并详细阐述了各个部分的组成及功能； 3、 介绍了定位参数计算过程中使用的图像二值化方法， 图像高斯平滑滤波方法， 图像分割方法和用于圆心坐标提取的广义 Hough 变换算法， 并对每个算法完成了相关实验； 4、 对于目标靶位姿计算问题， 采用矩阵奇异值分解和最小二乘法结合的方法完成了位姿参数计算； 5、 采用空间向量法作为计算模型， 最终完成车轮相应参数的计算。