Machine vision soap bubble method air tightness test method

Abstract

The invention discloses a machine vision soap bubble method air tightness test method. The method comprises the following steps: (1) gas filling, (2) liquid soap painting, (3) optical illumination, (4) photo shooting, (5) image preprocessing, (6) binarized image acquisition, (7) target data set making, (8) target image detection, (9) leakage point positioning and (10) leakage point repairing. According to the invention, images acquired by the high-pixel CCD camera are utilized, and accurate positioning and leakage rate estimation of leakage points are realized through leakage point image feature identification, so that the traditional manual air tightness inspection operation is replaced, the professional skill requirements on inspection personnel are reduced, the phenomena of missing inspection and false inspection are prevented, and the automation, digitization and intelligence of air tightness inspection are realized.

Description

technical field [0001] The invention belongs to the field of machine vision used in product quality inspection, specifically relates to artificial intelligence technologies such as machine vision and deep learning, and provides a machine vision airtightness inspection method. Background technique [0002] Airtightness is a key technical indicator for the normal operation of heat exchangers, refrigeration devices, pressure vessels and many other pressure and vacuum equipment that require airtightness. Existing air tightness test methods mainly include: differential pressure method, soap bubble method, pressure water test method, halogen method, helium mass spectrometry and nitrogen-hydrogen method, etc. Helium mass spectrometry and other methods can achieve high leak rate measurement accuracy, but "I only know the existence of leaks, but I don't know where the leaks are." [0003] Soap bubble method and pressure water detection method are by far the most traditional and comm...

AI Technical Summary

Problems solved by technology

[0007] 2. Bubbles are raised by the buoyancy of water, and can only stay on the surface of the leak point momentarily, easily fall off from the leak point, and move upward until the bubble bursts;

[0008] 3. When the bubbles are in the water, only deformed bubbles can be observed due to the refraction of the water;

[0009] 4. Impurities are easily mixed into the water, which increases the turbidity of the water; the bubbles interact with the impurities during the rising process, which will cause the bubbles to burst

[0010] The above-mentioned problems in the pressure water detection method will bring technical difficulties to the machine vision identification of air bubbles and location of leaks

Images