A method for detecting the size of the glass shell of a car lamp based on machine vision

Abstract

The invention proposes a machine vision-based method for measuring the size of a car lamp glass shell. The present invention is divided into four steps: firstly, correct the image of the currently acquired glass shell, and remove its horizontal inclination; secondly, use the Canny operator to perform edge detection, and on this basis, determine the vertical position of the glass shell in the image. The boundary position; thirdly, determine the horizontal boundary position of the glass shell in the image; finally, measure parameters such as length, outer diameter, and wall thickness. The invention has the advantages of not being affected by environmental factors, high robustness, fast detection speed, high precision and the like.

Description

technical field [0001] The invention relates to the field of machine vision, in particular to a machine vision-based method for detecting the geometric dimensions of a glass shell of a vehicle lamp. Background technique [0002] Automobile headlights are the key components of automobile lighting, and their quality affects the appearance and performance of the vehicle. Among them, the glass housing is an important component of the lamp. In the molding process of the glass shell of the car lamp, due to factors such as complex production process and many production links, it is inevitable that the geometric dimensions such as length, outer diameter, and thickness of the finished glass shell of the car light do not meet the standards, and these glass shells It cannot enter the subsequent car lamp manufacturing process, and must be inspected and sorted. [0003] The traditional method of manually measuring the parameters of the glass shell mainly relies on skilled technical wor...

AI Technical Summary

Problems solved by technology

However, this method has the following problems: (1) Since the mechanical device is transmitting the glass shell, it is difficult to keep the glass shell appearing in the same approximate area every time, and the a priori preset search area is difficult to be suitable for all situations ;(2) Since the bayonet of the conveyor belt is slightly larger than the glass shell, the glass shell may be deflected at a small angle in the image during the process of mechanically transporting the glass shell, which also brings difficulties to the visual measurement of the glass shell size; ( 3) The acquired image is greatly affected by the noise and light on the scene, and the gray distribution is uneven. With the passage of time of the detection equipment, the light source is prone to aging or the lens is polluted by dust. At this time, the acquired image will have Large changes, and the existing visual measurement method of glass shell size is greatly affected by environmental factors

Images