DLT algorithm needs at least 6 points for pose estimation from 3D-2D point correspondences. (expecte

DLT algorithm needs at least 6 points for pose estimation from 3D-2D point correspondences. (expected: 'count >= 6')

# -*- coding: utf-8 -*-
# 测试使用opencv中的函数solvepnp
import cv2
import numpy as np
# tag_size = 0.05
tag_size_half = 1110.025
fx = 610.32366943
fy = 610.5026245
cx = 313.3859558
cy = 237.2507269
K = np.array([[fx, 0, cx],[0, fy, cy],[0, 0, 1]], dtype=np.float64)
objPoints = np.array([[-tag_size_half, -tag_size_half, 0],[tag_size_half, -tag_size_half, 0],[tag_size_half, tag_size_half, 0],[-tag_size_half, tag_size_half, 0]], dtype=np.float64)
# imgPoints = np.array([[608, 167], [514, 167], [518, 69], [611, 71]], dtype=np.float64)objPoints = np.array([(-165.0, 170.0, -110.5),  # 左目左端(165.0, 170.0, -110.5),  # 右目右端,  # Left eye(0.0, 0.0, 0.0),  # Nose tip(-150.0, -150.0, -120.5),  # 口の左端(150.0, -150.0, -120.5)], dtype=np.float64)  # Right Mouth corner)imgPoints=np.array([[1030,528],
[1080,528],
[1050,541],
[1030,556],
[1057,556]],dtype=np.float64)cameraMatrix = Kdist_coeffs = np.array([0,0,0,0], dtype=np.double)
retval,rvec,tvec  = cv2.solvePnP(objPoints, imgPoints, cameraMatrix, distCoeffs=dist_coeffs)#,flags=cv2.SOLVEPNP_SQPNP)
# cv2.Rodrigues()
print(retval, rvec, tvec)

没有找到解

换一个求解算法：

# -*- coding: utf-8 -*-
# 测试使用opencv中的函数solvepnp
import cv2
import numpy as np
# tag_size = 0.05
tag_size_half = 1110.025
fx = 610.32366943
fy = 610.5026245
cx = 313.3859558
cy = 237.2507269
K = np.array([[fx, 0, cx],[0, fy, cy],[0, 0, 1]], dtype=np.float64)
objPoints = np.array([[-tag_size_half, -tag_size_half, 0],[tag_size_half, -tag_size_half, 0],[tag_size_half, tag_size_half, 0],[-tag_size_half, tag_size_half, 0]], dtype=np.float64)
# imgPoints = np.array([[608, 167], [514, 167], [518, 69], [611, 71]], dtype=np.float64)objPoints = np.array([(-165.0, 170.0, -110.5),  # 左目左端(165.0, 170.0, -110.5),  # 右目右端,  # Left eye(0.0, 0.0, 0.0),  # Nose tip(-150.0, -150.0, -120.5),  # 口の左端(150.0, -150.0, -120.5)], dtype=np.float64)  # Right Mouth corner)imgPoints=np.array([[1030,528],
[1080,528],
[1050,541],
[1030,556],
[1057,556]],dtype=np.float64)cameraMatrix = Kdist_coeffs = np.array([0,0,0,0], dtype=np.double)
retval,rvec,tvec  = cv2.solvePnP(objPoints, imgPoints, cameraMatrix, distCoeffs=dist_coeffs,flags=cv2.SOLVEPNP_SQPNP)
# cv2.Rodrigues()
print(retval, rvec, tvec)

DLT algorithm needs at least 6 points for pose estimation from 3D-2D point correspondences. (expecte相关推荐

cv::StereoCalibrate 源码解析（三）—— 部分函数解析
cvCalcMatMulDeriv: // reimplementation of dAB.m /*这个矩阵求导和传统的矩阵求导不太一样求dcda 假设a:M*N c:I*J 最后生成矩阵 IJ * ...
【CVPR2019】完整论文列表二
CVPR 2019 Paper list No.1001-1294 ?CVPR2019 完整列表一论文题目与链接 Semantic Component Decomposition for Face ...
ECCV 2018 完整论文集 -- List 下载链接
下文列表为ECCV2018官网得到了今年接收论文列表,共779篇: 下文为ECCV2018的全部接收论文汇总 Oral: Convolutional Networks with Adaptive Co ...
CVPR 2018 paper list（论文列表）
原文链接:http://cvpr2018.thecvf.com/program/main_conference 52 Embodied Question Answering Abhishek Das ...
基于视觉的机器人抓取: 论文及代码(Vision-based Robotic Grasping: Papers and Codes)
本文同步于微信公众号:3D视觉前沿,欢迎大家关注. 本文总结了基于视觉的机器人抓取的相关论文及代码,同步于 GitHub. 机器人抓取必需的信息是相机系下抓取器的6DoF位姿,包括抓取器的3D位置和抓 ...
PCL函数库摘要——3D点云特征描述与提取
1.Class pcl::Feature< PointInT, PointOutT > 类Feature是所有特征相关模块中其他类的基类,定义了所有描述子常用的函数接口. #include ...
CVPR 2018 paper ---object detection
转载:https://blog.csdn.net/qq_34848537/article/details/82968217 参考链接 CVPR 2018 论文解读集锦(9月27日更新) CVPR 20 ...
CVPR2019论文题目中文列表
英文题目中文题目 Finding Task-Relevant Features for Few-Shot Learning by Category Traversal 少镜头学习中用类别遍历法寻 ...
（九：2020.08.27）CVPR 2019 追踪之论文纲要（译）
CVPR 2019 追踪之论文纲要(修正于2020.08.28) 讲在前面论文目录讲在前面论坛很多博客都对论文做了总结和分类,但就医学领域而言,对这些论文的筛选信息显然需要更加精细的把控,所以自 ...

DLT algorithm needs at least 6 points for pose estimation from 3D-2D point correspondences. (expecte

DLT algorithm needs at least 6 points for pose estimation from 3D-2D point correspondences. (expecte相关推荐

最新文章

热门文章