Ultrasonic full-focusing image defect processing method and device based on kernel density estimation

The invention discloses an ultrasonic full-focus image defect processing method and device based on kernel density estimation. The method comprises the following steps: acquiring an ultrasonic full-focus image and an image matrix of the ultrasonic full-focus image, and performing image preprocessing on the ultrasonic full-focus image according to the image matrix to obtain a preprocessed image; obtaining an ultrasonic intensity threshold value and a kernel density estimation value of each pixel point based on the preprocessed image, and setting a pixel value corresponding to the pixel point according to a comparison result of the kernel density estimation value of each pixel point and the ultrasonic intensity threshold value to obtain an intermediate processed image; and binarizing the intermediate processing image to obtain a target image, and positioning and quantifying a defect area in the target image. Artifacts in the ultrasonic full-focus image can be effectively eliminated, the defects in the ultrasoni