This invention discloses an online detection method for
coating defects in
lithium-
ion battery electrodes based on
machine vision, belonging to the field of
lithium-
ion battery
manufacturing quality inspection technology. Addressing the problem of missed detection of minute metallic foreign objects by conventional dry film optical detection due to insufficient contrast in high-speed
coating lines, this invention deploys a piezoelectric
acoustic emission sensor array in pairs on the lip plate of the
coating die. Differential
processing suppresses background vibrations, captures
acoustic emission pulses from foreign objects, and extracts acoustic features to predict
foreign object parameters. Based on the predicted parameters, a thin-film interference
physical model is used to calculate the expected template for interference fringes in the downstream wet film region. The
acoustic emission event triggers a downstream camera to acquire interference images at the predicted position and extract features. The expected template is matched with the actual observation, and the
foreign object is confirmed, classified, and its embedding depth is evaluated through the acoustic-optical matching
score. This method improves the detection sensitivity of metallic foreign objects to the 40μm level at coating speeds above 80m / min, and can identify fully embedded foreign objects that cannot be detected by purely optical methods.