Infrared weak and small target detection method based on multi-scale feature enhancement

• Main Authors: XIANG QIANG, HU KUN, LU JINZHENG, LI QIANG
• Format: Patent
• Language: Chinese; English
• Published: 21.06.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

The invention discloses an infrared weak and small target detection method based on multi-scale feature enhancement. The infrared weak and small target detection method comprises the steps of S1, acquiring a single-frame infrared weak and small target data set based on a real scene; s2, preprocessing the single-frame infrared weak and small target data set to obtain a preprocessed data set; and S3, inputting the preprocessed data set into a YOLO-MSFE network, and carrying out infrared weak and small target detection by using the YOLO-MSFE network to obtain an infrared weak and small target detection result. The method can solve the problems that the existing infrared image has the characteristics of lack of color information, very sparse small targets, much noise interference and the like, and the detection effect is poor due to the large depth of the network and insufficient information utilization. 本发明公开了一种基于多尺度特征增强的红外弱小目标检测方法，包括：S1：获取基于真实场景的单帧红外弱小目标数据集；S2：对所述单帧红外弱小目标数据集进行预处理，得到预处理后的数据集；S3：将所述预处理后的数据集输入至YOL