视觉SLAM十四讲(第二版)ch11运行./feature_training和./loop_closure结果异常。ubuntu22.04

 ubuntu22.04在运行slambook2中的ch11代码,feature_training输出的结果多了许多数字(还不知道什么原因。。。)。并且loop_closure的结果非常不准确,image0 对应的评分最高(除了image0本身)的居然不是image9。

最终解决:在网上看有说是DBoW3库的问题的,也有说是Ubuntu系统版本的问题。我打开了我的ubuntu18.04(虚拟机),并且使用第一版代码中DBoW库重新安装。最终运行结果算是正常:

至少跟Image0相比评分最高的是image9了,至于为什么具体的评分和书上的不一样,我想是opencv版本不一样,提取的ORB特征有些许差别吧,我想。

如果有其他差错,请把build文件夹删了重新编译

### SLAM14 Course Chapter 7 Materials For the seventh chapter of the SLAM14 course, materials typically focus on advanced topics within Simultaneous Localization and Mapping (SLAM). This area explores methodologies to improve mapping accuracy and efficiency through various algorithms and techniques. In this context, while specific content may vary depending on the curriculum design, common themes include: - **Graph-Based SLAM**: Discusses how graph optimization can enhance map consistency over time[^1]. - **Loop Closure Detection**: Explores methods for recognizing previously visited locations to correct drift errors in maps. - **Multi-Robot SLAM Systems**: Investigates strategies where multiple robots collaborate to build a single coherent map more efficiently than individual units could achieve alone. To gain deeper insights into these subjects, resources like textbooks or academic papers covering machine learning applications in robotics would be beneficial. Additionally, practical coding exercises using libraries such as ROS (Robot Operating System) are often included to reinforce theoretical knowledge with hands-on experience[^3]. ```python import rospy from sensor_msgs.msg import LaserScan def scan_callback(msg): print(f"Received {len(msg.ranges)} laser ranges") rospy.init_node('laser_scan_subscriber') sub = rospy.Subscriber('/scan', LaserScan, scan_callback) rospy.spin() ``` This Python code snippet demonstrates subscribing to laser scanner data from a robot equipped with LIDAR sensors—a typical setup used in many SLAM implementations.
评论 6
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值