EmguCV编程第二天

今天,研究了各种不同的特征算子的性能。

发现:图像分辨率为640×480时,SIFT实时帧频可达到10FPS,SURF为15FPS,ORB为16FPS左右。

众所周知,SIFT速度最慢,但是感觉SURF速度不是SIFT的3倍么?好像没有这么快啊,另外,ORB效率应该是更高才对。所以我怀疑,应该是图像特征点绘制和显示耗时了。


然后,我又把特征点绘制部分代码注释掉,结果发现:竟然没有任何的影响!!实在是搞不懂。


今天又弄懂了一个事,在Emgu.CV 的最新版(3.1)中,SIFT和SURF虽然受到专利保护了,但是还是可以用的,也不需要再引入什么额外的库啥的。我估计唯一的问题是:如果想利用SIFT或者SURF来开发商业盈利的产品,那是不行了。对于其他的学习和普通的应用,还是没有什么限制的。


另外需要注意:BruteForceMatcher(暴力匹配)函数在新版中改名字了,改为了BFmatcher,看起来更简洁了!当然,我发现在Emgu.CV 的最新版(3.1)中,很多比较长的名字都被简化了!整体的安装包比以前的版本也小了很多很多!


下图是今天的收获,只实现了不同特征点的实时显示。明天接着实现各算子的实时匹配,看看到底哪种算子最稳定。


另外,还要提一下定时器的问题,

针对窗体的Timer 

            //优点一:线程安全;
            //优点二:一个Tick事件在前一个Tick事件被处理完毕前不会被触发(就是说它会耐心等待)。
            //优点三:你可以直接在Tick事件处理代码中更新控件。


这个里面的学问很大,一两句说不清楚。

Preface 1 Chapter 1: Introduction to Emgu CV 5 What is Emgu CV? 5 Comparing image-processing libraries 6 License agreement 6 Documentation and other material 7 Ease of use 7 Performance 8 Summary of the comparison 9 Advantages of Emgu CV 9 Cross-platform 9 Cross-language support with examples 9 Other advantages 10 Summary 10 Chapter 2: Installing Emgu CV 11 Downloading Emgu CV 11 Installing Emgu CV 11 Installing on Windows 11 Installing on Linux 16 Getting the dependency 16 Building Emgu CV from source 17 Installing on OS X 18 Getting the dependency 19 Building Emgu CV from source 19 Troubleshooting 20 Windows 20 Linux 21 OS X 22 Summary 22 Chapter 3: Hello World 23 Hello World in C# 23 Creating a new project 24 Designing our form 27 Coding 27 Output 29 Hello World in VB.NET 30 Hello World in C++ 31 Summary 32 Chapter 4: Wrapping OpenCV 33 Architecture overview 33 OpenCV 33 Emgu CV 34 Function mapping 36 Structure mapping 36 Enumeration mapping 37 Summary 37 Chapter 5: Working with Images 39 Digital image representation 39 Pixels and data 39 Pixel resolution 40 Color image representation 41 Color depth 43 Working with images 44 Creating an image 44 Loading an image from a file 46 Operations with pixels 47 Method naming rules 49 Using operator overload 50 Generic operations support 51 Garbage collection 51 XML serialization 52 Summary 53 Chapter 6: Working with Matrices 55 Matrix and the Image class 55 Definition and parameters 56 Working with matrices 56 Creating a matrix 57 Operations with elements 58 Summary 59 Chapter 7: Shape Detection 61 Canny Edge Detector 61 Hough transforms 63 Hough Line transform 63 Hough Circle transform 65 Contour 67 Contour finding 68 Representation of contours 68 Sequences of vertexes 68 Free chain codes 69 Drawing contours 69 Polygon approximations 70 A contours example 70 Summary 72 Chapter 8: Face Detection 73 Biometric systems 73 Camera captures 75 Machine learning 76 Face detection or the Haar classifier 77 Boosting theory and supervised learning 78 Haar-like features 78 Code for face detection 81 Summary 83 Chapter 9: License Plate Recognition 85 License Plate Recognition 85 Algorithms for LPR 86 OCR 87 Tesseract-OCR 88 Code for License Plate Recognition 88 Assumption 89 Source code 89 GetWhitePixelMask 90 DetectLicensePlate 91 FindLicensePlate 92 Output 93 Summary 93 Chapter 10: Image Stitching 95 Image stitching 95 Algorithms for image stitching 96 Image matching 96 Image calibration 97 Image blending 97 Code 98 Summary 99 Index 101
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