A one-dimensional calibration object consists of three or more collinear points with known relative positions. It is generally believed that a camera can be calibrated only when a 1D calibration object is in planar motion or rotates around a &mac

:研究了一种多目标识别算法,该算法用SUSAN角点形成SIFT特征点,采用阶梯图像金字塔结构实现尺度不变,为所有匹配点建立统一的超定线性方程组并对该方程组系数矩阵进行简 化使其维数降低一半,得到增广矩阵.对增广矩阵进行列变换,依据坐标转换的特性可从中提取多目标的稳定正常点,实现了快速分离多目标的匹配点. -: Study of a multi-target recognition algorithm using SUSAN corner formed SIFT feature point

按照二维函数的特点和视觉机制,提出了用来捕捉纹理基元的纹理检测器函数,基于纹理检测器和扩展的小波变换,提出了基于能量分解的影像纹理多尺度分析方法,并按照神经动力学的侧抑制和端点抑制等理论,实现了对多尺度纹理特征的融合，这一多尺度分析方法直接将影像纹理能量在时间一尺度空间分解,包含了相位信息,避免了基于线性变换多尺度分解引起的能量与相位分离,为纹理分析提供了一个层次性的框架,有效提高了纹理的识别能力。-According to the characteristics of two-dimensio

This paper describes the direction finding device that can be adapted to shipboard or mobile vehicles used for ES, ELINT and monitoring equipments. It has improved the limited capability factor of the linear array multi interferometer, the