Shulou(Shulou.com)06/01 Report--

This article shows you how to understand SLAM, the content is concise and easy to understand, can definitely brighten your eyes, through the detailed introduction of this article, I hope you can gain something.

"some of my friends are curious about what Xiaobai does. Here is a unified reply. Xiaobai is a student at school, and his current research direction is visual SLAM. Today, Xiaobai will give you a brief introduction to SLAM."

01

-

What is SLAM?

SLAM (Simultaneous Localization And Mapping), translated into Chinese as "simultaneous location and mapping". To put it simply, it is the technology of drawing a two-dimensional or three-dimensional map of your environment while determining your location. The more academic explanation is:

It means that the main body carrying a specific sensor establishes the model of the environment in the process of motion and estimates its own motion at the same time without prior information of the environment. -- Davison

Here is a quote from Davison to explain what SLAM is. Davision is the first SLAM project to realize monocular real-time online mapping. The MonoSLAM paper published in 2007 ranked second in the number of Caucus (cited 1152 times), 73 times less than the first (the first was published in 2006). It can be said that this article has a very important influence in the field of modern SLAM.

If a robot can move and avoid obstacles independently in an unknown environment, it must know where it is and the state of its environment. SLAM solves two key questions for mobile robots: where I am and the state of my surroundings. At the same time, because the SLAM technology can draw the scene map, it is also convenient for people to understand the environment which can not directly reach the area. Therefore, SLAM technology is widely used in unmanned vehicles, drones, AR, VR, logistics robots, floor sweeping robots and other fields.

02

-

Common classification

According to the different sensors used to obtain environmental information, it can be divided into laser SLAM (Lidar-SLAM) and visual SLAM (VSLAM). The overall idea of the two in the technical implementation is the same, but in the face of different data information, we need to adopt the corresponding solution, so there will be obvious differences.

Laser SLAM: lidar is widely used to obtain environmental information because of its high ranging accuracy, simple calculation process and high reliability. The driverless cars of Google, Uber and Baidu all use laser SLAM technology as the main environment sensing and positioning technology. Although the use of lidar is not limited by light conditions and can be used around the clock, this method is not suitable for underwater because water will affect the propagation of laser. And the high cost of lidar also limits the field of its use.

Visual SLAM: use cameras to obtain environmental information, and there are more detailed categories according to the type and number of cameras, such as monocular SLAM, binocular SLAM,RGBD-SLAM. Because the camera is cheap, it makes up for the shortcomings of lidar, and at the same time, the way of "seeing" is more in line with the way that human beings perceive the world. However, this method has a large amount of calculation, is easy to be limited by care conditions, and its use is limited in large scenes, so it is often used in indoor scenes. However, with the improvement of hardware performance, visual SLAM has also been rapid development and application.

Both kinds of SLAM have their own advantages and disadvantages. In actual use, they are often used in combination with other sensors, such as IMU,GPS and other devices.

03

-

This is SLAM.

To say a lot about the internal composition and implementation principle of SLAM, it is not without using video to illustrate what is SLAM intuition. So in order to let the friends know that "this is the SLAM", Xiaobai shows the dynamic pictures of two SLAM to facilitate everyone's understanding.

First of all, we will bring you SLAM, which is implemented by Google using cartographer technology. Here is the URL of the open source code: https://github.com/googlecartographer/cartographer.

Next, the Hong Kong University of Science and Technology's SLAM project, which combines monocular and IMU, will also bring you an open source link: https://github.com/HKUST-Aerial-Robotics/VINS-Mono.

The above content is how to understand SLAM. Have you learned any knowledge or skills? If you want to learn more skills or enrich your knowledge reserve, you are welcome to follow the industry information channel.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.