Since severe droughts are occurring more frequently and lengthening the dry season in the Amazon Rainforest, it is important to detect wildfires promptly and forecast possible spread for effective suppression response. Though computer vision researchers have applied algorithms to automatically detect wildfires, current models are computationally expensive and not versatile enough for the low technology conditions of South American wildfire hotspots. This comprehensive deep learning study first trains a Fully Convolutional Neural Network with skip connections on multispectral Landsat 8 images of Ecuador and the Galapagos. The model uses Green and Short-wave Infrared (SWIR) bands as inputs to predict each image's corresponding pixel-level binary fire mask. This model achieves a 0.962 validation F2 score and a 0.932 F2 score on test data from Guyana and Suriname. Afterward, image segmentation is conducted on the Cirrus band using K-Means Clustering to simplify continuous pixel values into three discrete classes representing differing degrees of cirrus cloud contamination. Two additional Convolutional Neural Networks are trained to classify the presence of a wildfire using these segmented cirrus images. The "experimental" model trained on the segmented inputs and SWIR data achieves a binary accuracy that is 2.306% higher than that of the "benchmark model" that is trained only on SWIR data. The difference in performance has a p-value of 0.00968. This proof of concept reveals that feature simplification can improve the performance of wildfire detection models. Overall, the software built in this study is useful for early and accurate detection of wildfires in South America.