Solid rocket engine defect automatic extraction method

According to the method, pixel points with structures similar to defects are separated through mathematical morphology low-hat transformation, and then the number of threshold values is determined by analyzing and judging a gray histogram of a separated image. And finally, extracting defects by adopting an OTSU threshold segmentation algorithm and opening and closing operation of mathematical morphology, wherein the method specifically comprises the following steps: preprocessing an original image; performing morphological low-hat transformation on the preprocessed image; drawing a gray level histogram of the image after low-hat transformation; judging the number of independent peaks in the gray histogram, and determining the category of image segmentation; according to a maximum between-cluster variance calculation method in an OTSU algorithm, as shown in a formula (1), calculating and drawing a variance map of the image; obtaining an optimal threshold value or threshold value combination according to the em