洛阳网站制作,苏州有名的互联网企业,wordpress 百度地图api,五八同城最新招聘信息这里写目录标题平滑#xff08;低通#xff09;空间滤波器盒式滤波器核平滑#xff08;低通#xff09;空间滤波器
平滑#xff08;也称平均#xff09;空间滤波器用于降低灰度的急剧过渡
在图像重取样之前平滑图像以减少混淆用于减少图像中无关细节平滑因灰度级数量不…
这里写目录标题平滑低通空间滤波器盒式滤波器核平滑低通空间滤波器
平滑也称平均空间滤波器用于降低灰度的急剧过渡
在图像重取样之前平滑图像以减少混淆用于减少图像中无关细节平滑因灰度级数量不足导致的图像中的伪轮廓平滑核与一幅图像的卷积会模糊图像
盒式滤波器核
盒式核是最简单的可分离低通滤波器其系数的值相同通常为1 m×nm\times{n}m×n的盒式滤波器为1的一个m×nm\times{n}m×n的阵列其前面有一个归一化的常数通过是1除以系数值之和通过是1/mn1/mn1/mn
def box_filter(image, kernel)::param image: input image:param kernel: input kernel:return: image after convolutionimg_h image.shape[0]img_w image.shape[1]m kernel.shape[0]n kernel.shape[1]# paddingpadding_h int((m -1)/2)padding_w int((n -1)/2)image_pad np.zeros((image.shape[0]padding_h*2, image.shape[1]padding_w*2), np.uint8)image_pad[padding_h:padding_himg_h, padding_w:padding_wimg_w] imageimage_convol image.copy()for i in range(padding_h, img_h padding_h):for j in range(padding_w, img_w padding_w):temp np.sum(image_pad[i-padding_h:ipadding_h1, j-padding_w:jpadding_w1] * kernel)image_convol[i - padding_h][j - padding_w] temp # 1/(m * n) * tempreturn image_convol# 盒式滤波器核
img cv2.imread(DIP_Figures/DIP3E_Original_Images_CH03/Fig0333(a)(test_pattern_blurring_orig).tif, 0)kernel_size [3, 11, 21]fig plt.figure(figsize(10, 10))
for i in range(len(kernel_size) 1):ax fig.add_subplot(2, 2, i1)if i 0:ax.imshow(img, cmapgray, vmin0, vmax255), ax.set_xticks([]), ax.set_yticks([])ax.set_title(Original)else:kernel np.ones([kernel_size[i-1], kernel_size[i-1]])kernel kernel / kernel.sizeimg_dst box_filter(img, kernel)ax.imshow(img_dst, cmapgray, vmin0, vmax255), ax.set_xticks([]), ax.set_yticks([])ax.set_title(fkernel size {kernel_size[i-1]})
plt.tight_layout()
plt.show()# 盒式滤波器核 可分离核的效果
img cv2.imread(DIP_Figures/DIP3E_Original_Images_CH03/Fig0333(a)(test_pattern_blurring_orig).tif, 0)kernel_size [3, 11, 21]fig plt.figure(figsize(10, 10))
for i in range(len(kernel_size) 1):ax fig.add_subplot(2, 2, i1)if i 0:ax.imshow(img, cmapgray, vmin0, vmax255), ax.set_xticks([]), ax.set_yticks([])ax.set_title(Original)else:kernel np.ones([kernel_size[i-1], kernel_size[i-1]])kernel kernel / kernel.sizeimg_dst separate_kernel_conv2D(img, kernel)img_dst normalize(img_dst) * 255ax.imshow(img_dst, cmapgray, vmin0, vmax255), ax.set_xticks([]), ax.set_yticks([])ax.set_title(fkernel size {kernel_size[i-1]})
plt.tight_layout()
plt.show()
