基于FPGA和DSP的红外图像预处理

编写分叉图的计算程序； 选择一个混沌区的控制参量a，并编写时域波形的计算程序(Bifurcation diagram of the preparation program Choose a zone of chaos control parameter a, and the preparation of time-domain waveform of the calculation program)

说明：  matlab软件进行近红外光谱图像处理mwpls方法，进行最佳光谱区间的选择(Matlab software for near infrared spectral image processing mwpls method, the best spectral range selection)

说明：  用小波变换对图像进行三级分解及求出各个子带系数(Using wavelet transform image decomposition and find three different sub-band coefficients)

对比BP，OMP，StOMP对自然图像的重构效果,输入图像是一幅256*256的lena图像，我们使用压缩采样后，再用上述三种方法对其进行重构，比较其结果，其中StOMP方法是用不同的阈值(Contrast to BP, OMP, StOMP effect on the reconstruction of natural images, input image is a 256* 256 lena image, we use the compressed sampled, and then reconstruct them the above-mentioned three methods to compare their results, which StOMP method is to use different thresholds)

Multivariate curve parsed - alternating least squares(Multivariate curve parsed - alternating further squares)

说明：  声呐或者雷达系统发射信号后，通过波束形成算法对两个目标（方位分别为theta1、theta2）回波进行仿真，经过相位补偿，可以看出图像有两个波峰，波峰处指向性函数峰值最大，峰值所对应的的角度值即为目标的方位角。所以该程序对数字波束形成（DBF）的原理进行了仿真，适合初学者掌握阵列信号处理中数字波束形成的原理。(After the sonar or radar system transmits the signal, the echo of the two targets (the azimuth is theta1 and theta2 respectively) is simulated by the beamforming algorithm. After the phase compensation, it can be seen that the image has two peaks, the peak of the directivity function is the largest, and the corresponding angle value of the peak is the azimuth of the target. So the program simulates the principle of DBF, which is suitable for beginners to master the principle of DBF in array signal processing.)

说明：  这个是在运行的时候为了调试而写的，主要应用的场合是循环多，跟踪不方便的时候，记录到日志，引用了一个单元Unit_HBase,Unit_LogSp，已上传(this is the time to run the debugger and writing, the main application is a revolving occasions, tracking inconvenient time to record log, quoted a unit Unit_HBase, Unit_LogSp, uploaded)

实现图像小波分解，可以自行指定分解级数，采用daubenchies小波(Achieve the wavelet decomposition, you can specify the decomposition of their own series, using wavelet daubenchies)

计算img文件夹内所有图像之间的相似度，并根据相似度排序，呈现每个图片的最相似图片结果(calculate the image similarity of each pic in img filedir)

说明：  实现图像配准，两幅不同来源的数据配准，可直接运行，效果直观(Image registration, two different sources of data registration)