The invention discloses a method for automatically identifying cracks in the whole
cracking process of high-compaction
expansive soil. The method comprises the steps of preparation of a soil body, building of a
drying model box, acquisition of crack images in the whole
cracking process and crack identification and quantification
processing. A crack image of the whole
cracking process of the soil body is obtained at a short time interval; according to the method, a Python automatic file reading and data exporting function is utilized, and OpenCV fracture contour identification is combined, so that batch identification and quantification of fracture soil
body images are realized, and data exporting and storage are automatically realized. The time, the
water content, the fracture rate, the fracture length, the fracture entropy and the like can be recorded at the same time in combination with the change of the
soil mass quality in the model box, and the fracture development condition of the whole
soil mass cracking process in the two dimensions of time and
water content is obtained. A crack image is analyzed by combining a PIV (particle image velocimetry) technology, micro-scale displacement of a soil body in the
drying process is captured, the cracking position of the soil body is accurately captured before cracking, and the stage of crack development is distinguished according to the displacement direction.