MultiScale Object Detection in Remote Sensing Images using Deep Learning

Abstract: With a rapid development in aerial technology, applications of Remote Sensing Images (RSI) have become more diverse. Remote sensing object detection is a difficult task due to complicated background, variations in the scales of the objects and proximity between objects of same scale. RSI’s are commonly captured from satellites with wide views, which leads to largescale images. The proposed model detects the objects at different scales. Feature Extraction and providing additional information about the object is done using Residual Neural Network101 (ResNet101) and ZFNet. Further, single scale and multiscale object detection is implemented using You Only Look Once (YOLOV5) and Faster Region based Convolutional Neural Network (Faster RCNN). A comparative study is done on all these techniques to evaluate the performance measures like Mean Average Precision and Accuracy.