Imaging Parameters-Considered Slender Target Detection in Optical Satellite Images

The existing slender target detection methods based on optical satellite images are greatly affected by the perspective and solar perspective. Due to limited data sources, it is difficult implement a fully data-driven approach. This work introduces imaging parameters of images, which reduces influence perspectives perspectives, demand for amount data. We improve oriented bounding box (OBB) detector faster R-CNN (region convolutional neural networks) propose an parameters-considered (IPC-Det) more suitable our task. Specifically, in first stage, umbra shadow extracted horizontal (HBB), respectively, then matching realized according parameters. In second paired features used complete classification regression, obtained OBB. experiments, after introducing parameters, accuracy improved 3.9% (up 87.5%), proving that this successful attempt introduce detection.