ORB检测

本文深入探讨ORB(Oriented FAST and Rotated BRIEF)特征检测器的原理和应用。首先,介绍了FAST关键点检测算法,它能快速找到图像中的角点。随后,讲解了BRIEF描述符如何生成紧凑且区分度高的二进制特征。接着,阐述了ORB如何改进这两者,包括关键点的方向赋值和旋转不变性。最后,讨论了ORB在实际应用中的优势,如在机器人定位、图像拼接和视觉SLAM等领域。

摘要生成于 C知道 ,由 DeepSeek-R1 满血版支持, 前往体验 >

import cv2 as cv
import os
path1=r'E:\sceneflow\driving\frames_finalpass\15mm_focallength\scene_backwards\fast\left\0001.png'
path2=r'E:\sceneflow\driving\frames_finalpass\15mm_focallength\scene_backwards\fast\right\0001.png'
img1=cv.imread(path1,1)
img2=cv.imread(path2,1)
orb = cv.ORB_create ()
keypoints1= orb.detect(img1, None)
keypoints2= orb.detect(img2, None)
keypoints1,desriptors1=orb.compute(img1,keypoints1)
keypoints2,desriptors2=orb.compute(img2,keypoints2)
print(desriptors1.shape,img1.shape)
print(desriptors2.shape,img2.shape)
outimg=None
outimg=cv.drawKeypoints(img1,keypoints1,outimg)
# cv.imshow("desriptor",outimg)

matcher=cv.BFMatcher()
matches=None
matches=matcher.match(desriptors1,desriptors2,matches)
min_dist=10000
max_dist=0
for i in range(len(desriptors1[0])):
    dist = matches[i].distance
    if dist < min_dist:
        min_dist = dist
    if dist > max_dist:
        max_dist = dist
print("max dist %f" % (max_dist))
print("min disr %f" % (min_dist))

good_matches=[]
for i in range(len(desriptors1[0])):
    if matches[i].distance<=max(2*min_dist,30.0):
        good_matches.append(matches[i])
img_match=None
img_goodmatch=None

img_match=cv.drawMatches(img1,keypoints1,img2,keypoints2,matches,img_match)
img_goodmatch=cv.drawMatches(img1,keypoints1,img2,keypoints2,good_matches,img_goodmatch)
cv.namedWindow("no optimize",cv.WINDOW_NORMAL)
cv.namedWindow("with optimize",cv.WINDOW_NORMAL)
cv.imshow("no optimize",img_match)
cv.imshow("with optimize",img_goodmatch)
cv.waitKey(0)
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值