Visual SLAM recent study, systematic study of the "visual SLAM fourteen From Theory to Practice 2nd Edition" e-book, also tested the code, visual SLAM very good introductory book, let me know the main part of the tree. Chapter 13 is particularly useful in the previous section can be a variety of small solutions to string together, to see a complete visual SLAM should look like.
"Visual SLAM fourteen From Theory to Practice 2nd Edition" e-book system introduced visual SLAM (simultaneous localization and mapping) required basic knowledge and core algorithms, both theoretical foundations of mathematics, such as rigid body motion in three-dimensional space, nonlinear optimization, and computer vision algorithms implemented including, for example, multi-view geometry, loop detection. In addition, it also provides a number of examples of code for study and research to more in-depth grasp of the content.
Version 2 adds more examples, added some experimental code to introduce the principle of the algorithm, more in-depth introduction to the underlying computing would be better, in addition to calling library functions, but also provides the underlying implementation. Lecture will speak 9 the first edition of the move to 13, after the introduction of all the necessary knowledge, to show you how to complete a SLAM system works, it will be implemented by a few hundred lines of code, complete front and rear ends SLAM system.
"Visual SLAM fourteen From Theory to Practice, Second Edition" e-book for entry slam is a good book, especially for white. I saw two or three months, from the depth of learning to slam see is this. Then you need further papers, this book has completed its mission.
In recent years, deep learning architecture due to the great success in computer vision applications and become popular. In particular convolutional neural network (CNN) deep learning has become the most important type of network structure. Learning computer vision depth study, design and deployment CNN, as well as basic knowledge of computer architecture visual depth.
"Convolution neural network and visual computing" e-books starting from zero base, the system elaborated convolution neural network theory and its practical application basis, can help to quickly learn and build depth learning system. It provides a wealth of theoretical knowledge and practical operation of the case, and a series of comprehensive toolkit to help in understanding the neural network and build convolution (CNN) necessary for basic information. It will focus on basic part convolution neural network, without involving concepts emerge in the advanced courses (CNN related topics).
"Convolution neural network and visual computing" e-book is divided into five chapters, Chapter 1 introduces the image representation and a number of computer vision models that are now being called manually modeling. Providing basic understanding of image representation, and introduces some linear and nonlinear representation or feature extraction method, and the characteristics of these representations. It describes some basic elements of an image (e.g., an edge) of the detection method further comprises the representation of some basic machine learning tasks.
Finishing "visual SLAM fourteen From Theory to Practice 2nd Edition" e-book information, "convolution neural network and visual computing" e-learning e-book data, computer vision for your reference:
HTTPS: // the WWW. yuque.com/mathpythondata/tr0r9u/ukh8kr
While watching e-books, while taking notes while debugging code, improve soon!