基于图像分割的红外图像撞击时间估计方法

Method of TTC estimation using image segmentation

以往的被动测距方法主要集中在可见光领域 ,利用图像稳定的特征点进行跟踪 ,目前利用计算目标尺度变化率进行撞击时间估计 ,已经发展成为一种较成熟的方法。然而在红外领域 ,由于红外图像自身的特性其特征点难以提取 ,原有的特征点跟踪方法无法适用。文中论述了一种红外图像撞击时间估计方法 ,根据红外图像难以提取特征点而相对易于目标分割的特点 ,提出了先对图像进行分割提取目标面积 ,然后根据图像序列目标面积变化率来进行撞击时间估计 ,并根据撞击时间系统特征对估计结果进行修正以进一步提高精度。

Most time\|to\|contact approaches were concentrated in visible light before, where the image feature points were stable, and can be traced between serial frames, so can time\|to\|contact estimation be obtained from object scale change rate. But in infrared, it is hard to get the stable feature which can be used to trace between images. In this paper, a new method of time\|to\|contact estimation is proposed. According to the characteristic of infrared image, which is hard to extract feature points but relative easier to segment, the OTSU segmentation method is developed to obtain the area of target, and then the interframe change ratio of the area is used to estimate the time\|to\|contact information. At last, an adaptive correction method is presented to increase the precision.