Question

正在尝试在python中实现任务here。我需要对灰度图像执行Garbor滤波，然后对garbor滤波图像执行高斯滤波。在观察Python代码和MATLAB代码的结果时，我发现它们完全不同。

在python中，有函数cv2.filter2D cv2.GaussianBlur用于Garbor和高斯滤波，就像MATLAB具有imgaborfilt和imgaussfilt一样。如代码片段所示，使用python函数将返回仅0和1的矩阵，而MATLAB将返回浮点数的矩阵。


def garbor_filters(wavelength, orientation):
  ksize = 39
  kern = cv2.getGaborKernel(ksize=(ksize, ksize), sigma = 0.5, theta=orientation, lambd=wavelength, gamma=math.sqrt(2), psi=0, ktype=cv2.CV_32F)
  return kern


def compute_garbor(img, filter):
 #filter: kernel obtained from the method above 
 filtered_img = cv2.filter2D(img, cv2.CV_8UC3, filter)
 return filtered_img

def getgaussianFiltrdImg(imageGarbors, filters):
  #filters: a tuple of wavelength and orientation  
  sigma = 0.5*filters[0]; #get wavelength
  K = 3;
  filtersize=2*math.ceil(2*K*sigma)+1
  gau_img = cv2.GaussianBlur(imageGarbors, (filtersize,filtersize), K*sigma)
  return gau_img


# call the methods
realimage = cv2.imread('kobi.png')
realimage = cv2.cvtColor(realimage, cv2.COLOR_BGR2GRAY)
img = rescale(realimage, 0.25, anti_aliasing=False)

imageGarbors=getImageGabor(img, (2.82,0))
imageGauss=getgaussianFiltrdImg(imageGarbors, (2.82,0))

Python在矩阵中仅返回0和1，而MATLAB返回连续值。我在做什么错了？

MATLAB的imgaborfilt和imgaussfilt函数的python版本是什么？

0 个答案: