Opencv learning notes (May 18) -- Reading the initial feeling of Master opencv

This series has not been updated for a long time. At the beginning of last December, I learned a new book "opencv 《Master opencv with practical computer vision ProjectsYou have never had time to read this book. You don't want to do any task before the Spring Festival. After reading it, I compared it with other books of opencv. The feeling is as follows:

Learning opencv is a classic old book. It is an entry-level teaching material. After reading it, you can know what opencv can do.CodeI personally think it is still a bit out, but the advantage is that there is a wide range of Chinese materials and many references can be found online.

Opencv cookbook is a book I have previously recommended. It was written based on version 2.2. I feel like I am using a textbook. The book aims to let everyone know how to call opencv functions and how to use opencv classes to implement simple visual tasks. Requires a certain degree of C ++ basics.

Master opencv is more like an upper-layer architecture, which is based on version 2.4. It teaches you how to use opencv to implement complex tasks and accomplish functions not provided by the built-in functions of opencv. (Because we often use "Is there XXX function in opencv ?" This book is to tell you that opencv is just a tool. It is hard to learn how to use this tool for secondary development)

The first chapter of this master opencv book is nothing complicated.

Chapter 1 introduces an edge detection, skin color detection, and fillingAlgorithmAnd port this function to the Android platform. The Edge Detection and filling algorithms are built-in functions of opencv. Here are some tips for reading the first chapter:

1. analyze specific tasks. Because the complexity of algorithms is put at the top of the embedded platform, bilateral filtering is simplified, and the padding algorithm is only calculated on a map with the source image reduced by 1/2. Skin Color Detection does not use the classic face detection algorithm. Instead, it uses a solution that is designed to run efficiently on an embedded platform. This method is worth learning.

2. The idea of object-oriented programming. This chapter also mentions the display of FPS, which is implemented by people in a class, and then compares the opencv learning notes () that I wrote before to display the current FPS and compete with the self-score. It's really self-defeating ~~.

3. It seems that you do not need to rewrite your code when porting your c ++ project to the Android platform. You can call C ++'sProgramIt feels a bit like Mex in MATLAB. The development of opencv in Android is basically incomprehensible.

Read this book over the past few days, and write down the Reading Notes for the last few chapters. I hope you can give me more comments.