Detection of Coconut Black Headed Caterpillar Using Aerial Imagery Combined with Artificial Intelligence

In this paper, the methods to detect coconut black headed caterpillars are investigated, which are presently difficult for locating and correcting in a timely manner. The infestation of coconut black headed caterpillars causes severe damage to the coconut trees resulting in reduction of yield significantly. The coconut trees, which are most facing this problem are mostly in the southern region, in Prachuap Khiri Khan Province and Surat Thani Province of Thailand that has the coconut trees, which are more than 20 meters high, making it impossible to observe and take care of it thoroughly. The developed drone can be used to inspect and find the abnormal area that may be caused by the infestation of coconut black headed caterpillars. The images from the drone will be processed by artificial intelligence system, when the flight is completed. These images are approved for the characteristic check by working with an entomologist, Department of Agriculture of Thailand, and additionally done for the ground check by placing the colour papers at the bottom of the coconut trees that are destroyed by coconut black headed caterpillars for labeling and for processing using the artificial intelligence system easily. RetinaNet model is used in this research for accuracy and precision. The result has 93% accuracy for rounds 3 of modeling.