Soil surfaces form complex crack networks as a result of water loss and shrinkage. A crack network destroys the integrity of the soil and becomes the main factor affecting rainfall infiltration, slope instability and soil integrity. In this paper, a soil fracture network is quantified using fractal characteristics and fractal dimensions, and the soil fracture network is identified and calculated using digital image processing technology. The fracture network of silty clay with different setaria viridis root content is studied during the process of evaporation. Saturated mud is prepared by taking soil samples and collecting setaria viridis roots. The content of setaria viridis roots in each saturated mud sample is 0 g, 0.1 g, 0.2 g and 0.3 g. In the artificial climate environment simulation system, thin-layer root soil is dried by controlling the temperature and humidity to simulate dry climate conditions. During the test, the crack development process is recorded using a digital camera. The results show that when the root content is 0, 0.1 g, 0.2 g and 0.3 g, the water content values when a fissure is generated are 35.2%, 41.2%, 42% and 46.4%, and the initial fractal dimension values are 1.100, 1.106, 1.112 and 1.115, respectively. The fractal dimension value increases rapidly in the early stage of fissure generation, and it reaches the maximum value when the water content reaches 13.66%, 15.2%, 15.66% and 17.98%, respectively. According to the change law of the fractal dimension, the fractal dimension increases rapidly following the initial appearance of the fracture, and, with a continuous reduction in water content in the later stage, the fracture characteristics gradually stabilize, and the change rate of the fractal dimension becomes slow.