Visual sensing algorithm for chrysanthemum cutting sticking robot system

摘要

The procedure for producing chrysanthemum transplants involves: making cuttings from mother plants, preserving the cuttings in a refrigerator until enough are ready for planting, and sticking them into a plug tray after applying water on them overnight. A commercial transplanter is used to transplant them in the field. A few lower leaves of the cutting should be removed when plants are ready for the transplanter, since the lower leaves would be obstacles for smooth transplanting, and this procedureis still manual. The robotic chrysanthemum cutting sticking system introduced consists mainly of a leaf removing device, a sticking device and a visual sensor. The visual sensor has to understand direction, location, size and shape of chrysanthemum cutting extraction features from an acquired image in the robotic cutting sticking system. A colour TV camera and a monochrome TV camera were used to discriminate target object from background. The monochrome TV camera, including infrared sensitivity, was effective at discriminating the cutting, because the cutting has much higher reflectance in the infrared region than in the visible region. After thresholding, the border between object and background was represented by chain codes using 8 values per image. The difference of chain code could indicate location of stem and its end of cutting. A manipulator could grip the cutting at an appropriate point and transport the cutting through the leaf removing device to the sticking device.