Determine_grid_rotation_fft.hdev相关例程学习

Index:…/Applications/Position-Recognition-2D/determine_grid_rotation_fft.hdev

compute_edge_direction_histogram (ImagePattern, ImageDirRef, HistogramDirRef)

开始先检测图片的边缘然后再把边缘方向直方图返回，顺便把数据存为一个数组。

plot_histo_scale (HistogramDirRef, WindowHisto, 'Angle [°]', 0, 359, 0, ValueYMax, 'black', 'blue', HistoBorder)

用之前计算的值绘制一个直方图在3601

NumImages := 7

for Index := 1 to 7 by 1

    read_image (Image, 'grid/grid_' + Index$'.2d')

    * 

    * FFT image of rotated object

    compute_edge_direction_histogram (Image, ImageTest, HistogramDirTest)

rft_generic (ImageTest, ImageFFTTest, 'to_freq', 'none', 'complex', 180)

接下来读取新的有旋转的图片，然后继续提取直方图等，并用直方图做rft变换

    * Compare FFT image of rotated object with reference FFT

    correlation_fft (ImageFFTRef, ImageFFTTest, ImageCorrelation)

    rft_generic (ImageCorrelation, ImageFFT, 'from_freq', 'none', 'real', 180)

    get_grayval (ImageFFT, Row, Column, Grayval)

create_funct_1d_array ([Grayval[179],Grayval,Grayval[0]], Function)

*

* Find best correlation between object and reference FFT

determine_max_correlation (Function, Angle)

寻找Angle这个函数里面有很多内容可以借鉴。