In view of the current problems such as the inability to detect the hole-forming property of millet and the lack of an accurate method for measuring the hole diameter, a detection method for the hole-forming property of millet based on the K-value method was proposed. Firstly, a criterion for the hole-forming property of millet based on the K-value method was established. Secondly, the boundary points of the sown millet were extracted based on image recognition, and a modified Welzl's algorithm was used to screen the key boundary points and calculate the hole diameter. Finally, a detection system for the hole-forming property of millet was designed, which can complete the shooting of the sown seeds and display the results of the circle-making effect, hole diameter, and hole-forming property detection on the detection interface. The experimental results showed that the time compression ratio of the modified Welzl's algorithm was greater than 1 compared with that of the Welzl's algorithm, RIA algorithm, and Dps algorithm, with the highest reaching 16.94. Moreover, the size of the minimum circumscribed circle made by the modified Welzl's algorithm was consistent with the results made by the Welzl's algorithm and RIA algorithm. The hole-forming property of the spoon chain seed delivery device for millet hole sowing under specific parameter settings was detected. The number of test groups with K values between 75% and 125% accounted for 82.6% of the total tests, the double-sowing situation accounted for 9.1%, and the missed-sowing situation accounted for 8.3%. This indicated that the seed delivery device had good seed sowing stability and hole-forming property under this parameter combination setting. The experimental results showed that this system had certain contributions to solving the practical problem of being unable to detect the hole-forming property.