GPU-based synthetic aperture radar target detection method

• Main Authors: QUAN HONGBIN, CUI ZONGYONG, CAO ZONGJIE
• Format: Patent
• Language: Chinese; English
• Published: 06.11.2018
• Subjects: CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention belongs to the technical field of radar remote sensing application and particularly relates to a GPU-based synthetic aperture radar target detection method. Based on the traditional two-parameter constant false alarm rate detection and convolutional neural network principles, the invention proposes a C-CFAR target detection method based on GPU parallel implementation, and the methodis applied to SAR target detection. By performing convolutional layer and pooling layer processing on an original image, performing two-parameter constant false alarm rate target detection and performing certain subsequent processing steps, a detection result is obtained, a GPU is used to carry out parallel acceleration on the algorithm, as a parallel implementation method, the speed of SAR targetdetection is greatly improved, and real-time requirements are satisfied. 本发明属于雷达遥感应用技术领域，具体涉及种基于GPU的合成孔径雷达目标检测方法。本发明基于传统的双参数恒虚警率检测和卷积神经网络原理，提出了种基于GPU并行实现的C-CFAR目标检测方法，并应用于SAR目标检测中。通过对原图像进行卷积层和池化层处理，然后进行双参数恒虚警率目标检测，并进行定的后续处理步骤