Design and implementation of IoT Based Helopeltis Sp. Control system in Cashew nut farm

Abstract

Agriculture is the main activity of the southern part of Tanzania residents, cashew nut being the most cash crop brought by Portuguese has never been yielding optimal yields due to the existence of pests especially Helopeltis sp. The Internet of Things (IoT) based Pest Control system aims to design the system which will be able to capture, identify and store the Helopeltis pest and also free the other non-targeted species. The system’s blueprint has been laid out through the Proteus and Google Colab (Collaboration) simulation tools. The pest recognition process has been done in the Google Colab Pro using tensorflow and other dependencies packages; the accuracy of 98.15% was obtained while the mechanical part of wiping pests into and out of the container has been facilitated using the Proteus. The last part of the farmer notification SMS (Short Message Service) has also been implemented using Proteus. The control of the cashew nut pests will now be in a green way by discouraging the use of pesticides which also destroy the pollinators and degrade the quality of the soil, the crop and the environment. It has been found that the image training requires adequate resources including graphic card, memory and processing power. Also it has been noted that the more the pictures are used for the training the more accuracy the detection will be.