...
首页> 外文期刊>Optical engineering >Real-time stitching method for infrared image
【24h】

Real-time stitching method for infrared image

机译:红外图像的实时拼接方法

获取原文
获取原文并翻译 | 示例
           

摘要

Infrared images acquired by forward-looking infrared camera have a limited field of view that is insufficient to monitor the entire scene with a single frame. To generate images containing the entire scene, infrared images can be assembled using image stitching. However, direct application of traditional image stitching methods has the disadvantage of long delays, large amounts of computation, and high memory storage. We present a real-time stitching method for infrared images. First, modified normalized cross correlation is introduced to find the overlapping area. Therefore, feature detection only needs to operate in overlapping areas. Also, scale-invariant feature transform is modified according to the properties of the infrared image. In addition, we implement the proposed method on an embedded system that is low cost, portable, and highly real-time. Our experimental results show the proposed method has almost the same quality of panorama compared with state-of-the-art methods, but the computation time is reduced significantly.
机译:由前视红外摄像机获取的红外图像的视野有限,不足以用单个帧监视整个场景。为了生成包含整个场景的图像,可以使用图像拼接来组合红外图像。然而,直接应用传统的图像拼接方法具有延迟长,计算量大和存储空间大的缺点。我们提出了一种红外图像的实时拼接方法。首先,引入修改的归一化互相关来找到重叠区域。因此,特征检测仅需要在重叠区域中进行操作。此外,比例不变特征变换根据红外图像的属性进行修改。另外,我们在低成本,可移植且高度实时的嵌入式系统上实现了所提出的方法。我们的实验结果表明,与最先进的方法相比,该方法的全景图质量几乎相同,但计算时间却大大减少了。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号