[ComputerVision]实验六：视差估计

目录

一、实验内容

二、实验过程

2.1.1  test.py文件

2.1.2  test.py文件结果与分析

2.2.1 文件代码

2.2.2  结果与分析

---

一、实验内容 给定左右相机图片，估算图片的视差/深度；体现极线校正（例如打印前后极线对）、同名点匹配（例如打印数量、或可视化部分匹配点）、估计结果（部分像素的视差或深度）。评估基线长短、不同场景（室内、室外）对算法的影响。 二、实验过程 2.1.1  test.py文件 from PIL import Image from pylab import * from scipy.ndimage import * import numpy as np import cv2 import matplotlib.pyplot as plt from scipy.ndimage import filters def plane_sweep_ncc(im_l, im_r, start, steps, wid):     m, n = im_l.shape     mean_l = np.zeros((m, n))     mean_r = np.zeros((m, n))     s = np.zeros((m, n))     s_l = np.zeros((m, n))     s_r = np.zeros((m, n))     dmaps = np.zeros((m, n, steps))     filters.uniform_filter(im_l, wid, mean_l)     filters.uniform_filter(im_r, wid, mean_r)     norm_l = im_l - mean_l     norm_r = im_r - mean_r     for displ in range(steps):         filters.uniform_filter(np.roll(norm_l, -displ - start) * norm_r, wid, s)         filters.uniform_filter(np.roll(norm_l, -displ - start) * np.roll(norm_l, -displ - start), wid, s_l)         filters.uniform_filter(norm_r * norm_r, wid, s_r)         with np.errstate(invalid='ignore'):             denominator = np.sqrt(s_l * s_r)             denominator[denominator == 0] = np.inf             dmaps[:, :, displ] = s / denominator     return np.argmax(dmaps, axis=2) def epipolar_correction(im_l, im_r, F):     h, w = im_l.shape     corrected_r = np.zeros_like(im_r)     for y in range(h):         for x in range(w):             pt = np.array([x, y, 1])             line = F @ pt             line = line / line[0]             a, b, c = line             u = int(round(-c / a))             v = int(round(-c / b))             if 0 <= u < w and 0 <= v < h:                 corrected_r[y, x] = im_r[v, u]                 print(f"\n校正前位置坐标： ({x}, {y}) -> 校正后位置坐标： ({u}, {v})")     return corrected_r def find_matches(im_l, im_r):     sift = cv2.SIFT_create()     kp1, des1 = sift.detectAndCompute(im_l.astype(np.uint8), None)     kp2, des2 = sift.detectAndCompute(im_r.astype(np.uint8), None)     bf = cv2.BFMatcher()     matches = bf.knnMatch(des1, des2, k=2)     good_matches = []     for m, n in matches:         if m.distance < 0.75 * n.distance:             good_matches.append(m)     return kp1, kp2, good_matches def compute_fundamental_matrix(kp1, kp2, matches):     points1 = np.float32([kp1[m.queryIdx].pt for m in matches])     points2 = np.float32([kp2[m.trainIdx].pt for m in matches])     F, mask = cv2.findFundamentalMat(points1, points2, cv2.FM_RANSAC)     return F def visualize_results(im_l, im_r, im_r_corrected, kp1, kp2, matches):     fig, axs = plt.subplots(1, 3, figsize=(15, 5))     axs[0].imshow(im_l, cmap='gray')     axs[0].set_title('Left Image')     axs[0].axis('off')         axs[1].imshow(im_r, cmap='gray')     axs[1].set_title('Right Image')     axs[1].axis('off')         axs[2].imshow(im_r_corrected, cmap='gray')     axs[2].set_title('Corrected Right Image')     axs[2].axis('off')         plt.show()         img_matches = cv2.drawMatches(im_l.astype(np.uint8), kp1, im_r.astype(np.uint8), kp2, matches[:10], None, flags=cv2.DrawMatchesFlags_NOT_DRAW_SINGLE_POINTS)     plt.figure(figsize=(10, 5))     plt.imshow(img_matches)     plt.title('Top 10 Matches')     plt.axis('off')     plt.show() im_l = np.array(Image.open('D:\\Computer vision\\KITTI2015_part\\left\\000000_10.png').convert('L'), 'f') im_r = np.array(Image.open('D:\\Computer vision\\KITTI2015_part\\right\\000000_10.png').convert('L'), 'f') steps = 50 start = 4 wid = 13 kp1, kp2, matches = find_matches(im_l, im_r) F = compute_fundamental_matrix(kp1, kp2, matches) im_r_corrected = epipolar_correction(im_l, im_r, F) visualize_results(im_l, im_r, im_r_corrected, kp1, kp2, matches) res = plane_sweep_ncc(im_l, im_r_corrected, start, steps, wid) imsave('D:\\Computer vision\\KITTI2015_part\\12_3test.jpg', res) 2.1.2  test.py文件结果与分析

上述代码通过特征点检测、基础矩阵计算、极线校正以及视差图计算实现了立体匹配和校正的流程。

结果一：

数据集如下图图1、图2所示，图3展示了极线校正前后坐标信息的部分截图，图4展示了部分同名点匹配结果，图5展示了视差估计结果。

图 1 left picture

图 2 right picture

图 3 极线校正前后坐标

图 4 同名点匹配图

图 5 视差估计结果

结果二：

数据集如下图图6、图7所示，图8展示了极线校正前后坐标信息的部分截图，图9展示了部分同名点匹配结果，图10展示了视差估计结果。

图 6 left picture

图 7 right picture

图 8 极线校正

图 9 同名点匹配

图 10 结果图 2.2.1 文件代码

a.stereo_module.py文件

from numpy import argmax, roll, sqrt, zeros from scipy.ndimage import filters def plane_sweep_ncc(im_l,im_r,start,steps,wid):     m,n=im_l.shape     mean_l=zeros((m,n))     mean_r=zeros((m,n))     s=zeros((m,n))     s_l=zeros((m,n))     s_r=zeros((m,n))         dmaps=zeros((m,n,steps))         filters.uniform_filter(im_l,wid,mean_l)     filters.uniform_filter(im_r,wid,mean_r)         norm_l=im_l-mean_l     norm_r=im_r-mean_r         for displ in range(steps):         filters.uniform_filter(roll(norm_l,-displ-start)*norm_r,wid,s)         filters.uniform_filter(roll(norm_l,-displ-start)*roll(norm_l,-displ-start),wid,s_l)         filters.uniform_filter(norm_r*norm_r,wid,s_r)                 dmaps[:,:,displ]=s/sqrt(s_l*s_r)     return argmax(dmaps,axis=2) def plane_sweep_gauss(im_l,im_r,start,steps,wid):     m,n = im_l.shape     # arrays to hold the different sums     mean_l = zeros((m,n))     mean_r = zeros((m,n))     s = zeros((m,n))     s_l = zeros((m,n))     s_r = zeros((m,n))     dmaps = zeros((m,n,steps))     filters.gaussian_filter(im_l,wid,0,mean_l)     filters.gaussian_filter(im_r,wid,0,mean_r)     norm_l = im_l - mean_l     norm_r = im_r - mean_r     for displ in range(steps):         filters.gaussian_filter(roll(norm_l,-displ-start)*norm_r,wid,0,s)         filters.gaussian_filter(roll(norm_l,-displ-start)*roll(norm_l,-displ-start),wid,0,s_l)         filters.gaussian_filter(norm_r*norm_r,wid,0,s_r)     dmaps[:,:,displ] = s/sqrt(s_l*s_r)     return argmax(dmaps,axis=2)

b. stereo_test.py文件

from matplotlib import colorbar from matplotlib.pyplot import imshow, show, subplot from numpy import array from PIL import Image import stereo_module as stereo import cv2 import matplotlib.pyplot as plt im_l=array(Image.open('D:\\Computer vision\\KITTI2015_part\\left\\000000_10.png').convert('L'),'f') im_r=array(Image.open('D:\Computer vision\\KITTI2015_part\\right\\000000_10.png').convert('L'),'f') steps=12 start=4 wid=9 res_ncc=stereo.plane_sweep_ncc(im_l,im_r,start,steps,wid) cv2.imwrite('D:\\Computer vision\\KITTI2015_part\\depth_ncc.png',res_ncc) res_gauss=stereo.plane_sweep_gauss(im_l,im_r,start,steps,wid) cv2.imwrite('D:\\Computer vision\\KITTI2015_part\\depth_gauss.png',res_gauss) subplot(121) imshow(im_l) subplot(122) imshow(res_ncc, cmap='jet') plt.colorbar() show() 2.2.2  结果与分析

视差估计结果如图11、图12所示

图 11 视差估计结果一

图 12 视差估计结果二