First of all! The most important thing is to master C++! Mastery refers to being very familiar with pointers, lists, vectors and other things, because these things are often used in image programming. Then there is the data structure based on C++. Commonly used are stack, linked list and binary tree, and there are various sorting algorithms. But the above things are recommended for you to learn by doing. They are not boring and easy to master.
Then linear algebra, geometry and vector mathematics. After laying the foundation of these courses (just pay attention to mastering the basic knowledge), you can learn the relevant knowledge of computer graphics (all graphics algorithms are written in C++, which is helpful for you to learn OpenGL in the future).
Finally, the study of OpenGL and various advanced algorithms ~
It usually takes more than one year for image programming and molding (not including the time to learn C++), so come on! It's actually not too difficult. Observe more at ordinary times and exercise your spatial imagination ~!
As for collision detection, it is actually another major. In the future, an image programmer will not be required to write collision detection. An engine is usually integrated by many classes. You just need to know the basic principles of collision detection.