Abstract: In the field of switchgear, the operating mechanism is the core component of the switch, with a wide variety of parts and complex shapes. Achieving fully automatic measurement of these parts dimensions is an essential requirement for intelligent manufacturing in this industry. A fully automatic measurement method and system of parts dimensions based on machine vision were studied. Firstly, the machine vision was used to acquire the image of the part in the bin, the center coordinate and attitude of the part contour were automatically identified, and the robot was used to grab the part from the bin to the loading station of multi-station measuring rotary turntable. Then, the automatic turntable rotated the inspected part, multiple industrial cameras were deployed at various inspection stations and installed on motion rails to automatically capture images of parts in all directions, and the dimensions of the part were measured through the image analysis algorithm. Finally, according to the measurement results, the inspected parts were placed into the corresponding material box by robot, thus completing the fully automatic measurement of complex shaped and multi-class parts. The test results show that the system measures the key dimensions of four types of parts and meets the measurement accuracy requirements for that part. And the system can automatically measure the dimensions of various complex shaped parts, with an average measurement time of 40 s for a single part.