the Vision based tuning of the control trajectory with JADE algorithm applied to leaves detection and cutting

Abstract

In the propagation of plant tissue in vitro, it is necessary to obtain circular sections of two square centimeters of area of the leaves. To achieve this type of cuts, visual feedback is used, for which the RGB planes are analyzed to detect the green color, and a set point to the trajectory controller is found as the centroid of the green image region. Tuning is done using the algorithms of the adaptive differential evolution of JADE, which allows reducing the position error to 0.001 centimeters. The processing time of the JADE algorithm found was 3604 seconds to determine the gains to satisfy the PD controller.