红外夜视观瞄系统的目标轮廓提取方法及FPGA实现

Edge detection algorithm for infrared night vision system and FPGA implementation

摘要:

针对红外夜视观瞄系统中的目标凸显需求，提出一种可用于红外图像的目标轮廓提取方法。该方法立足于实时红外成像系统，在分析红外目标图像特性的基础上，通过高斯滤波预处理、对比度增强以及暗背景边缘去除等操作，改善了Sobel算子在红外图像中应用的不足，仿真实现了红外图像目标边缘的准确提取。在现场可编程门阵列(FPGA)实时处理平台验证了算法在红外夜视观瞄系统中的应用效果。结果表明，针对红外夜视观瞄系统的目标轮廓提取方法在FPGA实时系统中具有良好的边缘提取效果，在占用低于2%的硬件资源条件下获得了较为显著的目标凸显性能。

Abstract:

In response to the need for target highlighting in infrared night vision aiming systems, a method for extracting target contours from infrared images is proposed. This method is based on real-time infrared imaging systems, and after analyzing the characteristics of infrared target images, the shortcomings of the application of Sobel operator in infrared images are improved through operations such as Gaussian filtering pre-processing, contrast enhancement, and dark background edge removal. The simulation has achieved accurate extraction of target edges in infrared images. The application effect of the algorithm in the infrared night vision aiming system is verified on a Field-Programmable Gate Array(FPGA) real-time processing platform. The results show that the method for extracting target contours in the infrared night vision aiming system has good edge extraction effects in the FPGA real-time system and has achieved significant target highlighting performance with less than 2% hardware resource occupation.