ABSTRACT

Deforestation through oil palm ranch development represents an agrarian drift with huge financial and natural impacts. From cleanser to donuts and ice cream, oil palm is present in numerous regular products, but many have never heard of it unequivocally. Since oil palm develops as it were in tropical situations, the crop’s extension has driven to deforestation, increased carbon outflows, and biodiversity misfortune, while at the same time giving numerous important occupations. With the financial employment of millions and the biological systems of the tropics at stake. This project presents an alternative method of working towards reasonable, opportune, and adaptable ways to address the development, administration and management of oil palm in Nigeria using High-resolution satellite imagery, As high-resolution imagery is a worldwide, regularly-updated, and exact source of information, when coupled with deep learning (computer vision algorithms), it presents a promising opportunity for computerized mapping of oil palm plantations, an important step toward understanding both local and worldwide impacts. This research study, designed and assessed a detection system, which uses a convolutional neural network to extricate important features, and a classifier trained using satellite images. Results show exceptional effectiveness with a training loss of 0.11 and an accuracy of 99.0%. Utilizing different images for validation taken from diverse elevations the model reached a training loss of 0.245 and an accuracy of 82.9999% on validation data, while on test data we have 1.59 in loss and an accuracy of 87.5%. Thus, the proposed approach is seemingly pertinent within the field of precision agriculture.