CPC G06T 7/215 (2017.01) [G01P 5/26 (2013.01); G06T 5/002 (2013.01); G06T 5/30 (2013.01); G06T 7/12 (2017.01); G06T 7/155 (2017.01); G06T 7/62 (2017.01); G06T 2207/20021 (2013.01); G06T 2207/20036 (2013.01); G06T 2207/30204 (2013.01); G06T 2207/30241 (2013.01); G06T 2207/30242 (2013.01)] | 8 Claims |
1. A method for morphological identification, trajectory tracking, and velocity measurement of a high-concentration microbubbles, comprising the following steps:
1) collecting high time-resolution initial images of microbubbles using a high-speed complementary metal-oxide semiconductor (CMOS) camera;
2) inputting a high time-resolution initial image of the microbubbles for pre-processing, where the pre-processing comprises binarization, noise removal, filling, and erosion of the high time-resolution initial image; and after the binarization, the gray level of a pixel is set to 1 in a bubble region and 0 in a non-bubble region, thereby providing a pre-processed image;
3) segmenting, based on a spatial distribution of the microbubbles, overlapped pixel blocks in the pre-processed image, thereby providing a pre-processed, segmented binary image;
4) identifying a boundary of each microbubble in the pre-processed, segmented binary image, obtaining morphological information of each of the microbubbles, performing morphological matching, calculating an equivalent diameter of all of the microbubbles in a microbubble group, and obtaining diameter distribution information, thereby providing a pre-processed, segmented and identified binary image; and
5) correlating images of the microbubbles at two adjacent instants, obtaining instantaneous velocity field information of the microbubbles, correcting a velocity vector of the obtained velocity field information, and displaying a motion trajectory and direction of each microbubble in the microbubble group,
wherein the operation to correlate images of the microbubbles at two adjacent instants of step 5), comprises:
correlating pre-processed, segmented, and identified binary images of the microbubbles at different instants; selecting a mass centroid of a pixel block of a certain microbubble in an image at a certain instant as a center, and extending a certain width and height in up-down and left-right directions respectively to form a pixel matrix, and denoting the pixel matrix as an interrogation window W1; determining, based on a position of the interrogation window W1 at a current instant, an interrogation window W with a same position, a same width, and a same height as the interrogation window W1 in an image at a next instant; determining all microbubbles comprised in the interrogation window W, and calculating a number of the microbubbles; denoting the number as N, and establishing N interrogation windows W2, each with a same width and height as the interrogation window W1, based on mass centroids of the N microbubbles to form N pixel matrices, where the number of the microbubbles is equal to a number of reference matrices; in the N interrogation windows W2, each interrogation window W2 corresponds to one reference matrix; correlating the pixel matrix corresponding to the interrogation window W1 at the current instant with N reference matrices at the next instant to obtain a correlation coefficient Rij; determining a reference matrix with a maximum correlation coefficient as a pixel region corresponding to a microbubble at the center of the interrogation window W1 at the current instant in an image of the microbubble at the next instant; and recording a correspondence between the pixel regions of the two microbubbles;
where the correlation coefficient Rij is calculated as follows:
where a numerator of the correlation coefficient is defined by a sum of pixel areas in an overlap region of the interrogation windows W1 and W2; B1 and B2 denote sums of areas of all pixel blocks comprised in the interrogation windows W1 and W2 respectively; i denotes a mark of the microbubble in the center of the interrogation window W1 at the current instant; and j denotes a mark of a microbubble in the center of a certain pixel matrix selected at the next instant.
|