Artificial Intelligence-Based Harmful Bird Deterrent System in Smart Agriculture

This research was primarily carried out to ensure effective application of artificial intelligence in the form of Convolutional Neural Network as bird deterrent solution in a rice farm in smart agriculture on real time. Here, the most specialized rice eater of all granivorous birds was identified as Java sparrow or Java rice bird known scientifically as Padda Oryzivora. In this research work, an enhanced Convolutional Neural Network (CNN) model was developed and applied in identifying the Padda Oryzivora with the training data sets as the different pictures of the birds. The different data collected as the different pictures of the bird as captured by the cameras were pre-processed by Raspberry pi 2 and used as the training data sets for this research in Google colab platform using python language. Then, an algorithm was developed to monitor bird classification by the developed enhanced convolutional neural network (CNN) and use such classification to either activate the speakers using the predator sound when harmful sparrow birds were classified by the model. Subsequently, the developed algorithm remained inactive when a beneficial bird was classified by the model. Once, the harmful sparrow birds (Padda Oryzivora) heard the sound of the predator (squirrel), it assumed that the squirrel was around and flew away. This work was simulated in google colab platform using 700 images of both harmful sparrow and beneficial birds. The developed model was able to show accuracy of 97%, precision of 96.5%, recall of 0.965 and f1 score of 0.965.The essence of this research work was to allow other beneficial birds like insectivorous ones which feed on insects and other plant pests to stay and feed on those pests for biological pest control which would increase the rice yield in the long run. It only activated the scaring sound of the predator (squirrel) when a sparrow bird was sighted by the cameras within the rice farm. Keywords: Artificial Intelligence, Smart Farming, Rice Sparrow Bird, Predator Sound and Machine Learning.