Global navigation method based on deep learning

The invention discloses a global navigation method based on deep learning. The method comprises the following steps: step 1, placing a part to be measured on a measuring platform of an image measuringinstrument; 2, shooting through a large-view-field camera to obtain a global image of the to-be-measured part; 3, the deep learning software identifies the type of the to-be-detected part through theshot global image; and 4, calling a measurement program of the corresponding part, and guiding a main image system of the image measurement instrument to a measurement area of the part to be measured. The invention has the advantages that: 1, the time for manually fixing the part to be measured on the clamp is greatly reduced, and the operation of manually creating and calling a measurement program is reduced, so that personal errors and insufficient efficiency are eliminated; and 2, the measuring equipment has higher expansibility, automatic positioning of the to-be-measured part is realized, a measuring program is a