Learning Computer Vision — An intro

The idea behind computer vision appears to be the process of using computing power to analyze images and discriminate between components in the image.
To give an HD image of a flower garden and ask a computer to tell me what different flowers there are, visible in the image, I would need to write algorithms to maybe:

Identify which of the different components seen in the picture are flowers. (Something that could potentially be implemented in numerous ways.

• How flowers are structured (with petals, etc)
• How flowers have starkly different colors from the rest of the components in the garden (maybe harder to implement due to how many exceptions seem to creep up because leaves, soil and bright sunlight also have starkly different colors and may be identified as flowers)

Discriminate between the different flowers. Maybe based on:

• How many petals there are
• The color of the flower
• The size of the flower
• The components constituting the flower

How many types of flowers there among all the flowers provided. (A simple .unique() method might do the job on this, but will still be drawing on an algorithm).