Count Leaves Detection of Hydroponic Pakcoy Using Region-based Convolutional Neural Networks Algorithm

Hydroponics is a farming technique where the planting medium does not use soil, the type of plant that is usually planted is a type of vegetable. This hydroponic technique is used because it does not require a large enough area to make it. However, monitoring the growth of hydroponic vegetables still relies on humans, so it requires time for routine monitoring directly. From this problem came the idea of creating a detection system for many Pakcoy hydroponic vegetable leaves to monitor the condition of hydroponic vegetables. The method used in this system is the R-CNN (Region-based Convolutional Neural Networks) method. The parameter used is counting many leaves from Pakcoy vegetables. This research count leaves of Pakcoy hydroponic using the R-CNN method obtained an accuracy of 86% with a system speed of 14.08 sec, partitioning data 10% testing data and 90% training data with a learning rate of 0,0002, epoch 6569, using an angle of 0 ° and a distance of 35 cm from Pakcoy vegetables to the webcam.