英语翻译The iterative scheme has been implemented and applied to image sequences corresponding to a number of simple flow patterns.The results shown here are for a relatively low resolution image of 32 by 32 picture cells.The brightness measureme
来源:学生作业帮助网 编辑:六六作业网 时间:2024/11/08 21:00:46
英语翻译The iterative scheme has been implemented and applied to image sequences corresponding to a number of simple flow patterns.The results shown here are for a relatively low resolution image of 32 by 32 picture cells.The brightness measureme
英语翻译
The iterative scheme has been implemented and applied to image sequences
corresponding to a number of simple flow patterns.The results shown here are
for a relatively low resolution image of 32 by 32 picture cells.The brightness
measurements were intentionally corrupted by approximately 1% noise and
then quantized into 256 levels to simulate a real imaging situation.The
underlying surface reflectance pattern was a linear combination of spatially
orthogonal sinusoids.Their wavelength was chosen to give reasonably strong
brightness gradients without leading to undersampling problems.Discon-
tinuities were avoided to ensure that the required derivatives exist everywhere.
Shown in Fig.5,for example,are four frames of a sequence of images
depicting a sphere rotating about an axis inclined towards the viewer.A
smoothly varying reflectance pattern is painted on the surface of the sphere.
The sphere is illuminated uniformly from all directions so that there is no
shading.We chose to work with synthetic image sequences so that we can
compare the results of the optical flow computation with the exact values
calculated using the transformation equations relating image coordinates to
coordinates on the underlying surface reflectance pattern.
英语翻译The iterative scheme has been implemented and applied to image sequences corresponding to a number of simple flow patterns.The results shown here are for a relatively low resolution image of 32 by 32 picture cells.The brightness measureme
迭代方案已经实施和应用相应的一些简单的流动图案的图像序列.
这里显示的结果是32个较低分辨率的图像32图像细胞.亮度测量是故意约1%噪声,然后量化为256级模拟实际成像情况损坏.底层表面的反射率模式空间正交的正弦信号的线性组合.他们的波长选择给相当强劲的亮度梯度没有导致欠采样问题.
断开tinuities避免了确保所需的衍生物无处不在.在图5所示,例如,在四帧的一个序列的描述一个球体的轴向观众旋转图像.一个平滑变化的反射图案画在球体的表面.
球被均匀地从各个方向,没有阴影.我们选择了与合成的图像序列,我们可以比较的光流计算结果与精确值的计算使用的图像坐标系中的坐标变换方程在底层表面的反射率模式.
希望采纳!不过,不采纳也没事的!
麻痹,这么多,懒着翻译