摘要
夜间低照度背景下,为了实现水炮对海面可疑目标的精确驱离,需要对水炮发射的射流轨迹进行提取。通过红外热像仪返回的红外图像中,射流轨迹、空气、海水、驱离目标等灰度信息互相干扰,因此在提取射流轨迹曲线前需要对图像进行基于频域分析的预处理滤波。对灰度图像逐行扫描并标记高频中心点,连接标记的相邻点成独立曲线,可以得到较为平滑的射流轨迹曲线与干扰曲线集合。对得到的曲线集合进行基于位置、方位角、轨迹长度、凹凸性、灰度等特征的筛选,确定满足射流轨迹特征的曲线并分析轨迹落点。实验结果表明,使用该图像处理方法提取射流轨迹的实时性和准确性都较为理想。
In order to achieve high precision when expelling suspicious target on the sea at low-light night, it＇s needed to extract the trajectory of jet launched from a water gun. In an infrared image which is obtained from the thermal infrared imager, the gray level of jet trajectory, air and sea water will interfere with each other, which results in an image pre-treated filter based on frequency domain before extracting the jet trajectory needed. The gray scale image is scanned line by line and high frequency center points are marked, then smooth curve set including jet trajectory curve and interference is got by linking a marked point to its neighbors. The curves are filtered based on features like position, azimuth, length of trajectory, convexity and concave, and gray level, the curve is selected which confirms with the features of jet trajectory, and the drop point of the selected curve is analyzed. The experiments show comparatively ideal real time and accuracy when the jet trajectory is extracted by using the above methods.
出处
《红外技术》
CSCD
北大核心
2017年第12期1120-1126,共7页
Infrared Technology
关键词
红外图像
射流轨迹
频域滤波
曲线提取
特征筛选
infrared image, jet trajectory, frequency filter, curve detection, feature-based filtering
