Abstract: There are the seeding cavities due to the low seeding number of super rice, which have influence on yield in rice. For solving the problem, seeding cavities of the super rice nursing tray were detected in continuous seeding process. The vision detection procedure was arranged between seeding and covering soil in rice seeding pipeline. The CCD(Charge Coupled Device) camera shot tray images continuously, and mask images were obtained which were consistent with the tray seeding cavities. The image information of storage buffer was achieved by reading program at certain time intervals. After the images were processed and analyzed, the position of seeding cavities were tracked and results were stored in seeding cavities database for re-seeding. The technology decreased cavities ratio of tray nursing seedlings and improved the seedling survival rate of super rice precision seeding.