SIFT原理與應用初探
導入需要用到的包
驗證旋轉不變性
import numpy as np
import cv2
import cv2 as cv
from matplotlib import pyplot as plt
使用cv2.xfeatures2d.SIFT_create()實例化SIFT函數,並且設置FLANN參數設計
sift = cv2.xfeatures2d.SIFT_create()
orb = cv.ORB_create()
# FLANN 參數設計
FLANN_INDEX_KDTREE = 0
index_params = dict(algorithm = FLANN_INDEX_KDTREE, trees = 5)
search_params = dict(checks=50)
flann = cv2.FlannBasedMatcher(index_params,search_params)
插入圖片1,並且進行灰度化處理,使用sift.detect(gray, None)生成關鍵點,使用sift.compute(kp) 求得關鍵點對應的特徵向量
img1 = cv2.imread('C:/users/Administrator/Desktop/1.jpg')
#使用cv2.imread()接口讀圖像,讀進來的是BGR格式以及[0~255]。所以要將img轉換爲RGB格式,不然後面顯示會有色差
img1 = cv2.cvtColor(img1,cv2.COLOR_BGR2RGB)
gray1 = cv2.cvtColor(img1, cv2.COLOR_BGR2GRAY) #灰度處理圖像
kp1, des1 = sift.detectAndCompute(img1,None)#des是描述符
kp, des = orb.detectAndCompute(img1,None)
print (des1.shape) #描述符數組維度
print(len(kp1)) #關鍵點個數
print(des.shape) #ORB
print(len(kp))
(1085, 128)
1085
(500, 32)
500
插入圖片2,並且進行灰度化處理,使用sift.detect(gray, None)生成關鍵點,使用sift.compute(kp) 求得關鍵點對應的特徵向量
img2 = cv2.imread('C:/users/Administrator/Desktop/2.jpg')
#使用cv2.imread()接口讀圖像,讀進來的是BGR格式以及[0~255]。所以要將img轉換爲RGB格式,不然後面顯示會有色差
img2 = cv2.cvtColor(img2,cv2.COLOR_BGR2RGB)
gray2 = cv2.cvtColor(img2, cv2.COLOR_BGR2GRAY)
kp2, des2 = sift.detectAndCompute(img2,None)
kpkp, desdes = orb.detectAndCompute(img2,None)
print (des2.shape) #描述符數組維度
print(len(kp2)) #關鍵點個數
(1089, 128)
1089
水平拼接,顯示原圖
print(img1.shape)
print(img2.shape)
hmerge = np.hstack((img1, img2)) #水平拼接
plt.figure(num=1,figsize=(16,16))
plt.imshow(hmerge)
plt.title('original')
plt.axis('off')
plt.show()
(400, 600, 3)
(400, 600, 3)
將圖片水平拼接並顯示
hmerge = np.hstack((gray1, gray2)) #水平拼接
plt.figure(num=1,figsize=(16,16))
plt.imshow(hmerge,cmap='gray')
plt.title('gray')
plt.axis('off')
plt.show()
使用cv2.drawKeypoints()進行畫圖操作,在圖中畫出關鍵點。
img3 = cv2.drawKeypoints(img1,kp1,img1,color=(255,0,255))
img4 = cv2.drawKeypoints(img2,kp2,img2,color=(255,0,255))
Img1 = cv2.drawKeypoints(img1,kp,img1,color=(255,0,255))
Img2 = cv2.drawKeypoints(img2,kpkp,img2,color=(255,0,255))
水平拼接並顯示
hmerge = np.hstack((Img1, Img2)) #水平拼接
plt.figure(num=1,figsize=(16,16))
plt.imshow(hmerge)
plt.title('keypoint')
plt.axis('off')
plt.show()
hmerge1 = np.hstack((Img1, Img2)) #水平拼接
plt.figure(num=1,figsize=(16,16))
plt.imshow(hmerge1)
plt.title('keypoint1')
plt.axis('off')
plt.show()
flann.knnMatch解決匹配並調整ratio
matches = flann.knnMatch(des1,des2,k=2)
matchesMask = [[0,0] for i in range(len(matches))]
good = []
for m,n in matches:
if m.distance < 0.7*n.distance:
good.append([m])
img5 = cv2.drawMatchesKnn(img1,kp1,img2,kp2,good,None,flags=2)
Img5 = cv2.drawMatchesKnn(img3,kp,img4,kpkp,good,None,flags=2)
顯示特徵匹配新效果圖
plt.figure(num=1,figsize=(16,16))
plt.imshow(img5)
plt.title('SIFT_rotation')
plt.axis('off')
plt.show()
plt.figure(num=1,figsize=(16,16))
plt.imshow(Img5)
plt.title('ORB_rotation')
plt.axis('off')
plt.show()
驗證尺度不變性
載入圖片並顯示尺寸大小和原始圖片
img1 = cv2.imread("C:/users/Administrator/Desktop/1.jpg")
img2 = cv2.imread("C:/users/Administrator/Desktop/3.jpg")
img1 = cv2.cvtColor(img1,cv2.COLOR_BGR2RGB)
img2 = cv2.cvtColor(img2,cv2.COLOR_BGR2RGB)
print(img1.shape)
print(img2.shape)
plt.figure(num=1,figsize=(12,12))
plt.imshow(img1,cmap='gray')
plt.title('original_img1')
plt.axis('off')
plt.show()
plt.figure(num=1,figsize=(12,12))
plt.imshow(img2,cmap='gray')
plt.title('original_img2')
plt.axis('off')
plt.show()
(400, 600, 3)
(200, 300, 3)
灰度化處理,打印描述符數組維度和關鍵點個數
gray1 = cv2.cvtColor(img1, cv2.COLOR_BGR2GRAY)
kp1, des1 = sift.detectAndCompute(img1,None)
Kp1, Des1 = orb.detectAndCompute(img1,None)
print (des1.shape) #描述符數組維度
print(len(kp1)) #關鍵點個數
(2336, 128)
2336
灰度化處理,打印描述符數組維度和關鍵點個數
gray2 = cv2.cvtColor(img2, cv2.COLOR_BGR2GRAY)
kp2, des2 = sift.detectAndCompute(img2,None)
Kp2, Des2 = sift.detectAndCompute(img2,None)
print (des1.shape) #描述符數組維度
print(len(kp2)) #關鍵點個數
(2336, 128)
770
顯示灰度化圖像
plt.figure(num=1,figsize=(12,12))
plt.imshow(gray1,cmap='gray')
plt.title('gray1')
plt.axis('off')
plt.show()
plt.figure(num=1,figsize=(12,12))
plt.imshow(gray2,cmap='gray')
plt.title('gray2')
plt.axis('off')
plt.show()
img3 = cv2.drawKeypoints(img1,kp1,img1,color=(255,0,255))
img4 = cv2.drawKeypoints(img2,kp2,img2,color=(255,0,255))
Img3 = cv2.drawKeypoints(img1,Kp1,img1,color=(255,0,255))
Img4 = cv2.drawKeypoints(img2,Kp2,img2,color=(255,0,255))
畫出關鍵點並顯示
plt.figure(num=1,figsize=(12,12))
plt.imshow(img3)
plt.title('keypoint_img3')
plt.axis('off')
plt.show()
plt.figure(num=1,figsize=(12,12))
plt.imshow(img4)
plt.title('keypoint_img4')
plt.axis('off')
plt.show()
plt.figure(num=1,figsize=(12,12))
plt.imshow(Img3)
plt.title('keypoint_img3')
plt.axis('off')
plt.show()
plt.figure(num=1,figsize=(12,12))
plt.imshow(Img4)
plt.title('keypoint_img4')
plt.axis('off')
plt.show()
matches = flann.knnMatch(des1,des2,k=2)
matchesMask = [[0,0] for i in range(len(matches))]
good = []
for m,n in matches:
if m.distance < 0.7*n.distance:
good.append([m])
img5 = cv2.drawMatchesKnn(img1,Kp1,img2,kp2,good,None,flags=2)
Img5 = cv2.drawMatchesKnn(img1,kp1,img2,Kp2,good,None,flags=2)
顯示特徵匹配效果圖
plt.figure(num=1,figsize=(16,16))
plt.imshow(img5)
plt.title('SIFT_size')
plt.axis('off')
plt.show()
plt.figure(num=1,figsize=(16,16))
plt.imshow(Img5)
plt.title('ORB_size')
plt.axis('off')
plt.show()
驗證亮度不變性
將原始圖片改變亮度和對比度,然後進行對比顯示
img1 = cv2.imread("C:/users/Administrator/Desktop/1.jpg")
img1 = cv2.cvtColor(img1,cv2.COLOR_BGR2RGB)
res = np.uint8(np.clip((1.5 * img1 + 10), 0, 255))
print(img1.shape)
print(res.shape)
tmp = np.hstack((img1, res)) # 兩張圖片橫向合併(便於對比顯示)
plt.figure(num=1,figsize=(16,16))
plt.imshow(tmp)
plt.title('original')
plt.axis('off')
plt.show()
(400, 600, 3)
(400, 600, 3)
gray1 = cv2.cvtColor(img1, cv2.COLOR_BGR2GRAY) #灰度處理圖像
kp1, des1 = sift.detectAndCompute(img1,None)#des是描述符
Kp1, Des1 = sift.detectAndCompute(img1,None)#des是描述符
print (des1.shape) #描述符數組維度
print(len(kp1)) #關鍵點個數
(1085, 128)
1085
gray2 = cv2.cvtColor(res, cv2.COLOR_BGR2GRAY) #灰度處理圖像
kp2, des2 = sift.detectAndCompute(res,None)#des是描述符
Kp2, Des2 = sift.detectAndCompute(res,None)#des是描述符
print (des2.shape) #描述符數組維度
print(len(kp2)) #關鍵點個數
(1066, 128)
1066
hmerge = np.hstack((gray1, gray2)) # 兩張圖片橫向合併(便於對比顯示)
plt.figure(num=1,figsize=(16,16))
plt.imshow(hmerge,cmap='gray')
plt.title('gray')
plt.axis('off')
plt.show()
img3 = cv2.drawKeypoints(img1,kp1,img1,color=(255,0,255))
img4 = cv2.drawKeypoints(res,kp2,res,color=(255,0,255))
Img3 = cv2.drawKeypoints(img1,Kp1,img1,color=(255,0,255))
Img4 = cv2.drawKeypoints(res,Kp2,res,color=(255,0,255))
hmerge = np.hstack((img3, img4)) #水平拼接
plt.figure(num=1,figsize=(16,16))
plt.imshow(hmerge)
plt.title('keypoint')
plt.axis('off')
plt.show()
hmerge = np.hstack((Img3, Img4)) #水平拼接
plt.figure(num=1,figsize=(16,16))
plt.imshow(hmerge)
plt.title('keypoint')
plt.axis('off')
plt.show()
flann.knnMatch解決匹配並調整ratio
matches = flann.knnMatch(des1,des2,k=2)
matchesMask = [[0,0] for i in range(len(matches))]
good = []
for m,n in matches:
if m.distance < 0.7*n.distance:
good.append([m])
img5 = cv2.drawMatchesKnn(img1,kp1,res,kp2,good,None,flags=2)
Img5 = cv2.drawMatchesKnn(img1,Kp1,res,Kp2,good,None,flags=2)
顯示特徵匹配效果圖
plt.figure(num=1,figsize=(16,16))
plt.imshow(img5)
plt.title('SIFT_brightness')
plt.axis('off')
plt.show()
plt.figure(num=1,figsize=(16,16))
plt.imshow(Img5)
plt.title('ORB_brightness')
plt.axis('off')
plt.show()
驗證仿射不變性
對原圖進行仿射變換並輸出
img = cv2.imread('C:/users/Administrator/Desktop/1.jpg')
img = cv2.cvtColor(img,cv2.COLOR_BGR2RGB)
rows, cols, ch = img.shape
pts1 = np.float32([[0, 0], [cols - 1, 0], [0, rows - 1]])
pts2 = np.float32([[cols * 0.2, rows * 0.1], [cols * 0.9, rows * 0.2], [cols * 0.1, rows * 0.9]])
M = cv2.getAffineTransform(pts1, pts2)
dst = cv2.warpAffine(img, M, (cols, rows))
plt.figure(num=1,figsize=(16,16))
plt.imshow(dst)
plt.title('affine_img')
plt.axis('off')
plt.show()
顯示原圖和仿射圖
print(img.shape)
print(dst.shape)
hmerge = np.hstack((img, dst)) # 兩張圖片橫向合併(便於對比顯示)
plt.figure(num=1,figsize=(16,16))
plt.imshow(hmerge)
plt.title('original')
plt.axis('off')
plt.show()
(400, 600, 3)
(400, 600, 3)
gray1 = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY) #灰度處理圖像
kp1, des1 = sift.detectAndCompute(img,None)#des是描述符
Kp1, Des1 = sift.detectAndCompute(img,None)#des是描述符
print (des1.shape) #描述符數組維度
print(len(kp1)) #關鍵點個數
(1085, 128)
1085
gray2 = cv2.cvtColor(dst, cv2.COLOR_BGR2GRAY) #灰度處理圖像
kp2, des2 = sift.detectAndCompute(dst,None)#des是描述符
Kp2, Des2 = sift.detectAndCompute(dst,None)#des是描述符
print (des2.shape) #描述符數組維度
print(len(kp2)) #關鍵點個數
(810, 128)
810
顯示灰度化圖像
hmerge = np.hstack((gray1, gray2)) # 兩張圖片橫向合併(便於對比顯示)
plt.figure(num=1,figsize=(16,16))
plt.imshow(hmerge,cmap='gray')
plt.title('gray')
plt.axis('off')
plt.show()
img3 = cv2.drawKeypoints(img,kp1,img,color=(255,0,255))
img4 = cv2.drawKeypoints(dst,kp2,dst,color=(255,0,255))
Img3 = cv2.drawKeypoints(img,Kp1,img,color=(255,0,255))
Img4 = cv2.drawKeypoints(dst,Kp2,dst,color=(255,0,255))
畫關鍵點並顯示
hmerge = np.hstack((img3, img4)) #水平拼接
plt.figure(num=1,figsize=(16,16))
plt.imshow(hmerge)
plt.title('keypoint')
plt.axis('off')
plt.show()
hmerge = np.hstack((Img3, Img4)) #水平拼接
plt.figure(num=1,figsize=(16,16))
plt.imshow(hmerge)
plt.title('keypoint')
plt.axis('off')
plt.show()
matches = flann.knnMatch(des1,des2,k=2)
matchesMask = [[0,0] for i in range(len(matches))]
good = []
for m,n in matches:
if m.distance < 0.7*n.distance:
good.append([m])
img5 = cv2.drawMatchesKnn(img,kp1,dst,kp2,good,None,flags=2)
Img5 = cv2.drawMatchesKnn(img,Kp1,dst,Kp2,good,None,flags=2)
print(img5)
print(Img5)
[[[255 255 255]
[255 255 255]
[255 255 255]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]
[[255 255 255]
[255 255 255]
[255 255 255]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]
[[255 255 255]
[255 255 255]
[255 255 255]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]
...
[[151 148 139]
[154 151 142]
[152 149 140]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]
[[140 140 132]
[143 143 135]
[142 142 134]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]
[[146 146 138]
[148 148 140]
[146 146 138]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]]
[[[255 255 255]
[255 255 255]
[255 255 255]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]
[[255 255 255]
[255 255 255]
[255 255 255]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]
[[255 255 255]
[255 255 255]
[255 255 255]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]
...
[[151 148 139]
[154 151 142]
[152 149 140]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]
[[140 140 132]
[143 143 135]
[142 142 134]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]
[[146 146 138]
[148 148 140]
[146 146 138]
...
[ 0 0 0]
[ 0 0 0]
[ 0 0 0]]]
顯示特徵匹配效果圖
plt.figure(num=1,figsize=(16,16))
plt.imshow(img5)
plt.title('SIFT_affine')
plt.axis('off')
plt.show()
plt.figure(num=1,figsize=(16,16))
plt.imshow(Img5)
plt.title('ORB_affine')
plt.axis('off')
plt.show()
orb 不支持放射不變性 orb提取特徵速度更快