Abstract: The present invention concerns a method for visually inspecting tubular garments such as socks, sleeves, stockings, panties or the like. The article is loaded onto or unloaded from a support or stretcher, and the article is checked or inspected by means of a computer-aided vision system during the loading onto or unloading off of the support or stretcher.

Abstract: Installation for the quality control of yarn bobbins, said installation comprising one or a plurality of quality control machines (2) which are associated to a system (5) for transporting the bobbins (7) onto supports (6), with stops for inspection purpose, said quality control machines having a housing (19) wherein are provided one or a plurality of inspection stations (11), one or a plurality of loading/unloading stations (20) and a unit for transferring the bobbins (7) from the transport system (5) to said housing (19), the bobbins being positioned successively in the stations (11, 20) and returning to the transport system (5); the unit for transferring the bobbins comprises a rotary frame (3) which makes stops of predetermined duration at each of the inspection stations (11) and loading/unloading stations (20) and incorporates various assemblies provided with a fixing member (4) intended to approach the support (6) carrying the bobbin (7) in order to take the later.

Abstract: The method which is applicable to the inspection of bobbins of yarn of any size, color and material, is based on the acquisition of images of the surface of the bobbin and on their processing by a computer program in order to evaluate defects, and comprises various steps of exploring the bobbin (1) taking a plurality of images of various areas of said bobbin, each step occurring under specific conditions of lighting and relative position of particular image sensors (3, 4) in order to identify specific defects, displacing either the bobbin (1) or the image acquisition device (3, 4), or bot to explore, sector by sector, the whole surface of interest, each of the focused images, related to a particular defect being processed individually by the evaluation program.