How can robots perceive the world and their own movements so that they accomplish navigation and manipulation tasks? In this module, we will study how images and videos acquired by cameras mounted on robots are transformed into representations like features and optical flow. Such 2D representations allow us then to extract 3D information about where the camera is and in which direction the robot moves. You will come to understand how grasping objects is facilitated by the computation of 3D posing of objects and navigation can be accomplished by visual odometry and landmark-based localization.

Voir plus

- Week 1 -
**Geometry of Image Formation**

Welcome to Robotics: Perception! We will begin this course with a tutorial on the standard camera models used in computer vision. These models allow us to understand, in a geometric fashion, how light from a scene enters a camera and projects onto a 2D image. ... - Week 2 -
**Projective Transformations**

Now that we have a good camera model, we will explore the geometry of perspective projections in depth. We will find that this projection is the cause of the main challenge in perception, as we lose a dimension that we can no longer directly observe. In this ... - Week 3 -
**Pose Estimation**

In this module we will be learning about feature extraction and pose estimation from two images. We will learn how to find the most salient parts of an image and track them across multiple frames (i.e. in a video sequence). We will then learn how to use featur... - Week 4 -
**Multi-View Geometry**

Now we will use what we learned from two view geometry and extend it to sequences of images, such as a video. We will explain the fundamental geometric constraints between point features in images, the Epipolar constraint, and learn how to use it to extract th...

**Kostas Daniilidis**, Professor of Computer and Information Science

School of Engineering and Applied Science**Jianbo Shi**, Professor of Computer and Information Science

School of Engineering and Applied Science

Coursera est une entreprise numérique proposant des formation en ligne ouverte à tous fondée par les professeurs d'informatique Andrew Ng et Daphne Koller de l'université Stanford, située à Mountain View, Californie.

Ce qui la différencie le plus des autres plateformes MOOC, c'est qu'elle travaille qu'avec les meilleures universités et organisations mondiales et diffuse leurs contenus sur le web.

Lire les avis

Trier par :

5 / 5

Terminé

Great Deal of Math.Prof. Shi's lectures on math guides me through this course. Whenever he shows up in the video, I know he will give me almost everything I need to solve the problems.Really Intensive and rewarding.The programming assignment is not that difficult if we have understood the meaning of the equations on the slide.But the math is not easy. Though Prof. Shi has been giving the lectures in a rather reasonable pace, I still have to pause the videos for quite a long time to follow him on math. I WILL NEVER FORGET SVD AFTER THIS COURSE. AMAZING!Hope Coursera can offer more intensive courses like this. Really like courses going in the order of advanced math - algorithm - practice.

0

5 / 5

Terminé

This is a Coursera course with the richest contents I ever had. Very glad to have learned so much in robotic perceptions. Thanks so much to Prof Daniilidis and Prof Shi. it is challenging but also very useful and helpful for further study or research. TAs are also very good helping lots of students. Love this class. Thank you all!

0

3 / 5

Terminé

The course is excellent is computer vision! The only problem it is not didactic at all, so if you don't are familiar with this content it will be very hard (even impossible) to follow.

0

5 / 5

Terminé

This course is interesting and very thorough. Some concepts of robot perception are explained in detail, with a focus on perception based on 2D vision. The videos are clear and there is a great number of quizzes and Matlab programming to improve your practical understanding of the topic. Be warned, though, that this course takes longer than 4 weeks in fact due to the numerous and long lectures.

0

5 / 5

Terminé

The course is very important for any student / engineer working in the field of robotics. It gives a lot of detailed information about the background needed as well as some hands-on experience with the basic tools in computer vision. A very good point is connecting what we study in the course with some real applications.

0

3 / 5

Terminé

some interesting material. The Slides for week 2 and 4 are terrible, too condensed with very little explanation on difficult topics. The Homeworks are pretty interesting, the assignments for week 3 and 4 complement eachother very well. the week 2 Kalman filter assignment didn't seem to work. I submitted something in frustration and was very surprised that it was accepted.

0

5 / 5

Terminé

Course is unusually difficult compared to the others in the series. You'll learn plenty of stuff, though, which is useful not just in robotics itself but many other applications with a mobile camera (such as stitching panoramas taken with your phone, or producing CGI).

0

4 / 5

Terminé

The content is not very easy to understand because the lecture speaks very fast and the document is not very sufficient. But in all, the content is good, help me with my research.

0

5 / 5

Terminé

This course is definitely worth learning if you are interested in computer vision or robotics perceptions! There are some minor flaws in the lectures slides, but it doesn't seriously effect the learning experience. I would recommend this course to people who have some basic knowledge about computer vision (e.g. camera calibration, coordinate transformation, affine/rigid transforms, linear solution of structure from motion). Otherwise, the latter part of this course could be a bit difficult.

0

3 / 5

Terminé

This course could use some help. It's a very interesting and important topic and is also difficult, but it could be explained better and the tie in between the lecture videos, quizzes and homework assignments could also be better. Some of the quiz questions are not answerable from reviewing the lecture notes and require outside knowledge of linear algebra and rotation mathematics. The assignments should also be better defined and set up so that there is incremental feedback available for the intermediate steps. For example, the last week's assignment has 5 steps, each of which requires a Matlab function to be written. In many online courses, there are "correct" intermediate results given so that each step can be verified before proceeding to the next step. In this assignment, there is not much feedback until you get to the third or fourth step and even then it's not the best. I had an error in one of the functions, but the problem feedback (photo comparisons) showed it as being OK until I submitted it for grading. It's important, since there's no instructor feedback , to provide some means of checking if you're doing things correctly.Some of the terminology used would be more clear if it was standardized; sometimes coordinates are x and y, sometimes u and v, there's also u1, u2, u3 and things like X = [x,y,z,w] and x = [u,v,w]. Its often quite difficult to know what's being referred to it's called x. I did learn a lot from this course, but it could have been a lot easier.

0

vous pourriez aussi être intéressé par...

MESSAGE_PLACEHOLDER