OpenCV_12 图像拼接

---

1.图像合并步骤

• 读文件并重置尺寸
• 根据特征点和计算描述子，得到单应性矩阵
• 图像变换
• 图像拼接并输出图像

2.自行编写的图像拼接程序

import cv2
import numpy as np 


# 特征点过滤阈值
verify_ratio = 0.8
# 过滤后特征点存储空间
verify_matches = []
min_matches = 8 # 最少特征点匹配数

# 获取单应性矩阵函数
def get_homo(img1, img2):
    # 创建特征转换对象
    sift  = cv2.SIFT_create()
    # 通过特征转换对象获得特征点和描述子
    k1, d1 = sift.detectAndCompute(img1, None)
    k2, d2 = sift.detectAndCompute(img2, None)
    # 创建特征匹配器,进行特征匹配
    bf = cv2.BFMatcher()
    matches = bf.knnMatch(d1, d2, k = 2)

    #过滤特征，找出有效的特征匹配点
    for m1, m2 in matches:
        if m1.distance < verify_ratio * m2.distance:
            verify_matches.append(m1)

    # 获取单应性矩阵
    if len(verify_matches) > min_matches:
        img1_pts = []
        img2_pts = []

        for m in verify_matches:
            img1_pts.append(k1[m.queryIdx].pt)
            img2_pts.append(k2[m.trainIdx].pt)

        img1_pts = np.float32(img1_pts).reshape(-1, 1, 2)  #修改数据格式，n行1列，每列2个数据
        img2_pts = np.float32(img2_pts).reshape(-1, 1, 2)

        H, mask = cv2.findHomography(img1_pts, img2_pts, cv2.RANSAC, 5)
        return H
    else:
        print('err: Not enough matches!')
        exit()
    
# 图像拼接函数
def stitch_img(img1, img2, H):
    # 获得图像的四个角点

    h1, w1 = img1.shape[:2]
    h2, w2 = img2.shape[:2]
    img1_dims = np.float32([[0, 0], [0, h1], [w1, h1], [w1, 0]]).reshape(-1, 1, 2)
    img2_dims = np.float32([[0, 0], [0, h2], [w2, h1], [w2, 0]]).reshape(-1, 1, 2)
   
    # 对图像进行变换（单应性矩阵试图进行旋转，平移）
    img1_transform = cv2.perspectiveTransform(img1_dims, H)
    print(img1_dims)
    print(img2_dims)
    print(img1_transform)
    # 创建一张大图，将两张图拼到一起
    result_dims = np.concatenate((img2_dims, img1_transform), axis = 0)
    print(result_dims)

    [x_min, y_min] = np.int32(result_dims.min(axis = 0).ravel() - 0.5)   # 返回每一列（x轴）最小值
    [x_max, y_max] = np.int32(result_dims.max(axis = 0).ravel() + 0.5)
    
    transform_dist = [-x_min, -y_min]   # 平移距离
    '''平移矩阵:
    [1, 0, dx
     0, 1, dy
      0, 0, 1]
    '''
    transform_array = np.array([[1, 0, transform_dist[0]],
                                                                    [0, 1, transform_dist[1]],
                                                                    [0, 0, 1]])
    result_img = cv2.warpPerspective(img1, transform_array.dot(H), (x_max - x_min, y_max - y_min))
    # 拼接图像
    result_img[transform_dist[1]:transform_dist[1]+ h2, 
                            transform_dist[0]:transform_dist[0]+ w2] =img2

    return result_img

# 读取文件，将图片设置为一样大小 640x480
img1 = cv2.imread('./picture/shan2.jpeg')
img2 = cv2.imread('./picture/shan1.jpeg')

img1 = cv2.resize(img1, (640, 480))
img2 = cv2.resize(img2, (640, 480))

#将两张图横向排列并显示
inputs = np.hstack((img1, img2))


# 找特征点，根据特征点和计算描述子，得到单应性矩阵
H = get_homo(img1, img2)
# 根据单应性矩阵进行图像变换，然后进行平移
result_img = stitch_img(img1, img2, H)



# 图像拼接并输出图像
cv2.imshow('result', result_img)
cv2.waitKey(0)

结果展示：

3.使用OpenCV提供的API：Stitcher_create

import cv2


# 读取文件，将图片设置为一样大小 640x480
img1 = cv2.imread('./picture/shan2.jpeg')
img2 = cv2.imread('./picture/shan1.jpeg')

img1 = cv2.resize(img1, (640, 480))
img2 = cv2.resize(img2, (640, 480))

imgs = []    

imgs.append(img1)
imgs.append(img2)
            
# 这一行需要注意
stitcher = cv2.Stitcher_create(cv2.Stitcher_PANORAMA)
status, result_image = stitcher.stitch(imgs)
if status != cv2.Stitcher_OK:
    print("Can't stitch images, error code = %d" % status)
    exit()
cv2.imshow('result', result_image)

cv2.waitKey(0)

结果展示：

---