This invention discloses a thickness measurement method based on gradient-weighted
image segmentation, belonging to the field of optical
interferometry technology. First, the upper surface of the attenuation-corrected
optical coherence tomography image is aligned, and the lower surface region to be measured is selected. A positive depth gradient and a lateral constraint gradient are constructed. A total gradient map is generated using gradient weighting combined with image
grayscale value weights. Then, a neighborhood local constraint search interval is established based on the position of the previous column A scan interface. The maximum gradient value of each interval is searched column by column. Finally, the
boundary contour is smoothed by
moving average filtering, and the
optical depth of the lower surface is obtained by solving for the global mean. The difference between the upper and lower surface optical depths is divided by the
refractive index to obtain the medium thickness. This invention enhances the gradient response of the
boundary region by adding
grayscale value weights to the attenuation-corrected image and suppresses point
noise interference using neighborhood constraints, thereby improving the accuracy of weak boundary thickness measurement.