1.提取图像描述符并保存文件中
新建代码generate_descriptors.py
import os
import cv2
import numpy as np
from os import walk
from os.path import join
import sys
def create_descriptors(folder):
files = []
for (dirpath, dirnames, filenames) in walk(folder):
files.extend(filenames)
for f in files:
save_descriptor(folder, f, cv2.xfeatures2d.SIFT_create())
def save_descriptor(folder, image_path, feature_detector):
print( "读取中... %s" % image_path)
if image_path.endswith("npy"):
return
img = cv2.imread(join(folder, image_path), 0)
keypoints, descriptors = feature_detector.detectAndCompute(img, None)
descriptor_file = image_path.replace("jpg", "npy")
np.save(join(folder, descriptor_file), descriptors)
dir = sys.argv[1]
create_descriptors(dir)
数据集采用100张panda图片;数据集已放入百度云。
链接:https://pan.baidu.com/s/1M3EQ0-44h1S2eoNhTs5qSg
提取码:po73
脚本执行时需要传递该panda图像文件夹;
我的panda文件夹和代码在同一路径:
python generate_descriptors.py panda
结果下如图,特征提取后保存在文件夹中:
如遇到SIFT算法版本问题,需要重装opencv-python==3.4.2.16;opencv-contrib-python==3.4.2.16这两个版本
安装过程见上一篇。https://mp.csdn.net/postedit/95511133
2.扫描匹配
将测试图片test.jpg与训练图片放在同一文件夹(panda)内
新建scan_for_matches.py
from os.path import join
from os import walk
import numpy as np
import cv2
from sys import argv
from matplotlib import pyplot as plt
# create an array of filenames
folder = argv[1]
query = cv2.imread(join(folder, "test.jpg"), 0)
# create files, images, descriptors globals
files = []
images = []
descriptors = []
for (dirpath, dirnames, filenames) in walk(folder):
files.extend(filenames)
for f in files:
if f.endswith("npy") and f != "test.npy":
descriptors.append(f)
print (descriptors)
# create the sift detector
sift = cv2.xfeatures2d.SIFT_create()
query_kp, query_ds = sift.detectAndCompute(query, None)
# create FLANN matcher
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)
# minimum number of matches
MIN_MATCH_COUNT = 10
potential_culprits = {}
print (">> 初始化扫描图片...")
for d in descriptors:
print( "--------- 分析 %s 匹配 ------------" % d)
matches = flann.knnMatch(query_ds, np.load(join(folder, d)), k =2)
good = []
for m,n in matches:
if m.distance < 0.7*n.distance:
good.append(m)
if len(good) > MIN_MATCH_COUNT:
print ("%s 可匹配! (%d)" % (d, len(good)))
else:
print ("%s 不匹配" % d)
potential_culprits[d] = len(good)
max_matches = None
potential_suspect = None
for culprit, matches in potential_culprits.items():
if max_matches == None or matches > max_matches:
max_matches = matches
potential_suspect = culprit
print ("可能匹配的是 %s" % potential_suspect.replace("npy", "").upper())
#与测试图片匹配的图片culprit_image
culprit_image = potential_suspect.replace("npy", "jpg")
target = cv2.imread(join(folder, culprit_image), 0)
target_kp, target_ds = sift.detectAndCompute(target, None)
#添加文字说明
font = cv2.FONT_HERSHEY_SIMPLEX
cv2.putText(target,'pipeijieguo.jpg ',(10,30), font, 1,(255,0,0),2,cv2.LINE_AA)
cv2.putText(query,'test.jpg',(10,30), font, 1,(255,0,0),2,cv2.LINE_AA)
cv2.imshow('pipeijieguo',target)
'''#01暴力匹配BFMatcher
bf = cv2.BFMatcher(cv2.NORM_HAMMING, crossCheck=True)
matches = bf.match(query_ds, target_ds)
matches = sorted(matches, key=lambda x: x.distance)
resultImage = cv2.drawMatches(query, query_kp, target, target_kp, matches[:80], target, flags=2)
'''
#03 获取flann匹配器
matches_o = flann.knnMatch(query_ds, target_ds, k =2)
well= []
for k,l in matches_o:
if k.distance<0.7*l.distance:
well.append(k)
if len(well)>MIN_MATCH_COUNT:
src_pts = np.float32([ query_kp[k.queryIdx].pt for k in well ]).reshape(-1,1,2)
dst_pts = np.float32([ target_kp[k.trainIdx].pt for k in well ]).reshape(-1,1,2)
M, mask = cv2.findHomography(src_pts, dst_pts, cv2.RANSAC,5.0)
matchesMask = mask.ravel().tolist()
h,w = query.shape
pts = np.float32([ [0,0],[0,h-1],[w-1,h-1],[w-1,0] ]).reshape(-1,1,2)
dst = cv2.perspectiveTransform(pts,M)
image = cv2.polylines(target,[np.int32(dst)],True,255,3,cv2.LINE_AA)
else:
print( "匹配数量不足 - %d/%d" % (len(well),MIN_MATCH_COUNT))
matchesMask = None
drawParams=dict(matchColor=(0,255,0),singlePointColor=None,matchesMask=matchesMask,flags=2)
resultImage=cv2.drawMatches(query,query_kp,target,target_kp,well,None,**drawParams)
plt.imshow(resultImage)
plt.show()
执行脚本
python scan_for_matches.py panda
使用test.jpg与数据集匹配结果如下图: