Keypoint matching - OpenCV 키포인트 매칭

Computer Vision 2019. 11. 2. 23:48 |
반응형


Keypoint matching.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
#include <opencv2/opencv.hpp>
 
using namespace std;
using namespace cv;
 
int main(int argc, char** argv)
{
    Mat trg = imread("Target.jpg", IMREAD_GRAYSCALE);
    Mat src = imread("Source.jpg", IMREAD_GRAYSCALE);
    
    if (trg.empty() || src.empty()) {
        cerr << "Image load failed!" << endl;
        
        return -1;
    }
 
    Ptr<Feature2D> feature = ORB::create();
    
    vector<KeyPoint> trgKeypoints, srcKeypoints;
    Mat trgDesc, srcDesc;
    feature->detectAndCompute(trg, Mat(), trgKeypoints, trgDesc);
    feature->detectAndCompute(src, Mat(), srcKeypoints, srcDesc);
    // Detects keypoints and computes the descriptors.
 
    Ptr<DescriptorMatcher> matcher = BFMatcher::create(NORM_HAMMING);
    // Brute-force matcher create method.
 
    vector<DMatch> matches;
    // Class for matching keypoint descriptors.
    matcher->match(trgDesc, srcDesc, matches);
    // Finds the best match for each descriptor from a query set.
 
    sort(matches.begin(), matches.end());
    vector<DMatch> good_matches(matches.begin(), matches.begin() + 100);
 
    Mat dst;
    drawMatches(trg, trgKeypoints, src, srcKeypoints, good_matches, dst,
        Scalar::all(-1), Scalar(-1), vector<char>(), DrawMatchesFlags::NOT_DRAW_SINGLE_POINTS);
    // Draws the found matches of keypoints from two images.
 
    imshow("dst", dst);
 
    waitKey();
    
    return 0;
}




100 best matches.


반응형
저작자표시 비영리 변경금지 (새창열림)

'Computer Vision' 카테고리의 다른 글

Simple copy and save in cv::imshow() 간단하게 이미지 복사, 저장 하기  (0) 2019.11.17
Finding a homography matrix - OpenCV 두 평면 사이의 투영 변환 행렬  (0) 2019.11.03
Detect, compute and draw keypoints - OpenCV 키포인트 찾기  (0) 2019.11.02
Screen capture - Windows API, OpenCV로 화면 캡쳐하기  (0) 2019.10.28
Windows dialog-based OpenCV program 윈도우 대화상자 OpenCV 프로그램 만들기  (2) 2019.10.26
Posted by J-sean
:

Detect, compute and draw keypoints - OpenCV 키포인트 찾기

Computer Vision 2019. 11. 2. 22:51 |
반응형

Detect, compute and draw key points.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
#include <opencv2/opencv.hpp>
 
using namespace std;
using namespace cv;
 
int main(int argc, char** argv)
{
    Mat src = imread("Source.jpg", IMREAD_GRAYSCALE);
    
    if (src.empty()) {
        cerr << "Image load failed!" << endl;
        
        return -1;
    }
 
    Ptr<Feature2D> feature = ORB::create();
    // static Ptr<ORB> cv::ORB::create(int nfeatures = 500, float scaleFactor = 1.2f, int nlevels = 8,
    // int edgeThreshold = 31, int firstLevel = 0, int WTA_K = 2, ORB::ScoreType scoreType = ORB::HARRIS_SCORE,
    // int patchSize = 31, int fastThreshold = 20)
 
    vector<KeyPoint> keypoints;
    feature->detect(src, keypoints);
    // Detects keypoints in an image (first variant) or image set (second variant).
 
    Mat desc;
    feature->compute(src, keypoints, desc);
    // Computes the descriptors for a set of keypoints detected in an image (first variant) or image set (second variant).
 
    cout << "keypoints.size(): " << keypoints.size() << endl;
    cout << "desc.size(): " << desc.size() << endl;
 
    Mat dst;
    drawKeypoints(src, keypoints, dst, Scalar::all(-1), DrawMatchesFlags::DRAW_RICH_KEYPOINTS);
    // Draws keypoints.
 
    imshow("src", src);
    imshow("dst", dst);
 
    waitKey();
    
    return 0;
}




Sizes of keypoints and descriptors.


Source image.


Keypoints with sizes and orientations.


반응형
저작자표시 비영리 변경금지 (새창열림)

'Computer Vision' 카테고리의 다른 글

Finding a homography matrix - OpenCV 두 평면 사이의 투영 변환 행렬  (0) 2019.11.03
Keypoint matching - OpenCV 키포인트 매칭  (0) 2019.11.02
Screen capture - Windows API, OpenCV로 화면 캡쳐하기  (0) 2019.10.28
Windows dialog-based OpenCV program 윈도우 대화상자 OpenCV 프로그램 만들기  (2) 2019.10.26
Convert Mat to Bitmap for Windows applications with GDI+ and surface crack detection - OpenCV Mat에서 비트맵으로 변환과 표면 크랙 검출  (0) 2019.10.26
Posted by J-sean
:

Screen capture - Windows API, OpenCV로 화면 캡쳐하기

Computer Vision 2019. 10. 28. 19:57 |
반응형

Screen capture with Windows API and OpenCV.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
#include <Windows.h>
#include <iostream>
#include "opencv2/imgproc.hpp"
#include "opencv2/highgui.hpp"
 
using namespace std;
using namespace cv;
 
class hWnd2Mat
{
public:
    hWnd2Mat(HWND hWindow, float scale = 1);
    virtual ~hWnd2Mat();
    virtual void Read();
    Mat capture;
 
private:
    HWND hWnd;
    HDC hWindowDC, hWindowCompatibleDC;
    int height, width, srcHeight, srcWidth;
    HBITMAP hBitmap;
    BITMAPINFOHEADER bi;
};
 
hWnd2Mat::hWnd2Mat(HWND hWindow, float scale)
{
    hWnd = hWindow;
    hWindowDC = GetDC(hWnd);
    hWindowCompatibleDC = CreateCompatibleDC(hWindowDC);
    SetStretchBltMode(hWindowCompatibleDC, COLORONCOLOR);
 
    RECT windowsize;    // get the height and width of the screen
    GetClientRect(hWnd, &windowsize);
 
    srcHeight = windowsize.bottom;
    srcWidth = windowsize.right;
    height = (int)(windowsize.bottom * scale);
    width = (int)(windowsize.right * scale);
 
    capture.create(height, width, CV_8UC4);
 
    // create a bitmap
    hBitmap = CreateCompatibleBitmap(hWindowDC, width, height);
    bi.biSize = sizeof(BITMAPINFOHEADER);    // http://msdn.microsoft.com/en-us/library/windows/window/dd183402%28v=vs.85%29.aspx
    bi.biWidth = width;
    bi.biHeight = -height;  //this is the line that makes it draw upside down or not
    bi.biPlanes = 1;
    bi.biBitCount = 32;
    bi.biCompression = BI_RGB;
    bi.biSizeImage = 0;
    bi.biXPelsPerMeter = 0;
    bi.biYPelsPerMeter = 0;
    bi.biClrUsed = 0;
    bi.biClrImportant = 0;
 
    // use the previously created device context with the bitmap
    SelectObject(hWindowCompatibleDC, hBitmap);
};
 
void hWnd2Mat::Read()
{
    // copy from the window device context to the bitmap device context
    StretchBlt(hWindowCompatibleDC, 0, 0, width, height, hWindowDC, 0, 0, srcWidth, srcHeight, SRCCOPY);
    //change SRCCOPY to NOTSRCCOPY for wacky colors!
    GetDIBits(hWindowCompatibleDC, hBitmap, 0, height, capture.data, (BITMAPINFO*)&bi, DIB_RGB_COLORS);
    //copy from hWindowCompatibleDC to hBitmap
};
 
hWnd2Mat::~hWnd2Mat()
{
    DeleteObject(hBitmap);
    DeleteDC(hWindowCompatibleDC);
    ReleaseDC(hWnd, hWindowDC);
};
 
int main()
{
    HWND hWndDesktop = GetDesktopWindow();
    hWnd2Mat desktop(hWndDesktop, 1);    // scale = 1
 
    cout << "Screen capure in 3 seconds." << endl;
    
    for (int i = 3; i > 0; i--)
    {
        cout << i << ".." << endl;
        Sleep(1000);
    }
 
    desktop.Read();
    imshow("Capture", desktop.capture);
 
    waitKey();
 
    return 0;
}



It captures your desktop image in 3 seconds.


반응형
저작자표시 비영리 변경금지 (새창열림)

'Computer Vision' 카테고리의 다른 글

Keypoint matching - OpenCV 키포인트 매칭  (0) 2019.11.02
Detect, compute and draw keypoints - OpenCV 키포인트 찾기  (0) 2019.11.02
Windows dialog-based OpenCV program 윈도우 대화상자 OpenCV 프로그램 만들기  (2) 2019.10.26
Convert Mat to Bitmap for Windows applications with GDI+ and surface crack detection - OpenCV Mat에서 비트맵으로 변환과 표면 크랙 검출  (0) 2019.10.26
Windows dialog based OpenCV Perspective Transform - 윈도우 다이얼로그 기반의 OpenCV Perspective Transform  (0) 2019.10.20
Posted by J-sean
:

Windows dialog-based OpenCV program 윈도우 대화상자 OpenCV 프로그램 만들기

Computer Vision 2019. 10. 26. 18:32 |
반응형

Build a Windows dialog-based OpenCV program.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
#include <Windows.h>
#include <gdiplus.h>
#include <opencv2/opencv.hpp>
#include "resource.h"
 
#pragma comment(lib, "gdiplus")
 
using namespace cv;
using namespace Gdiplus;
 
INT_PTR MainDlgProc(HWND hDlg, UINT iMessage, WPARAM wParam, LPARAM lParam);
HWND hDlgMain;
 
VideoCapture cap;
Mat frame;
RECT invalidateRect;


Prepare necessities.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, LPSTR lpszCmdParam, int nCmdShow)
{
    ULONG_PTR gpToken;
    GdiplusStartupInput gpsi;
 
    if (GdiplusStartup(&gpToken, &gpsi, NULL) != Ok) {
        MessageBox(NULL, TEXT("GDI+ start-up error."), TEXT("GDI+ Error"), MB_OK);
    }
 
    DialogBox(hInstance, MAKEINTRESOURCE(IDD_DIALOG1), HWND_DESKTOP, MainDlgProc);
    
    GdiplusShutdown(gpToken);
 
    return 0;
}


Start GDI+ and create a modal dialog box from a dialog template resource.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
bool initCamera()
{
    cap.open(0);
    if (cap.isOpened()) {
        invalidateRect = { 0, 0, cvRound(cap.get(CAP_PROP_FRAME_WIDTH)), cvRound(cap.get(CAP_PROP_FRAME_HEIGHT)) };
 
        return true;
    }
    else {
        MessageBox(NULL, TEXT("Camera open failed."), TEXT("Camera Error"), MB_OK);
 
        return false;
    }    
}


Initiate a camera and set the windows update region.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
void OnPaint(HDC hdc)
{
    Graphics G(hdc);
 
    cap >> frame;
    if (frame.empty()) {
        TextOut(hdc, 300, 240, "No frame to show", 8);
 
        return;
    }
 
    if (IsDlgButtonChecked(hDlgMain, IDC_CHECK1) == BST_CHECKED) {
        cvtColor(frame, frame, COLOR_BGR2GRAY);
        Canny(frame, frame, 50, 100);
        cvtColor(frame, frame, COLOR_GRAY2BGRA);
    }
    else {
        cvtColor(frame, frame, COLOR_BGR2BGRA);
    }
 
    Bitmap bitmap(frame.size().width, frame.size().height, frame.step1(), PixelFormat32bppARGB, frame.data);
    G.DrawImage(&bitmap, 0, 0, bitmap.GetWidth(), bitmap.GetHeight());
}


Process each frame and draw it. 


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
INT_PTR MainDlgProc(HWND hDlg, UINT iMessage, WPARAM wParam, LPARAM lParam)
{
    HDC hdc;
    PAINTSTRUCT ps;
    static HANDLE hTimer;
 
    switch (iMessage) {
    case WM_INITDIALOG:
        hDlgMain = hDlg;
        if (initCamera()) {
            hTimer = (HANDLE)SetTimer(hDlg, 1, 10, NULL);
        }
        else {
            hTimer = NULL;
        }
 
        return TRUE;
 
    case WM_COMMAND:
        switch (LOWORD(wParam)) {
        case IDOK:
 
            return TRUE;
 
        case IDCANCEL:
            if (hTimer != NULL) {
                KillTimer(hDlg, 1);
            }
            EndDialog(hDlg, IDCANCEL);
 
            return TRUE;
        }
 
    case WM_PAINT:
        hdc = BeginPaint(hDlg, &ps);
        OnPaint(hdc);
        EndPaint(hDlg, &ps);
 
        return TRUE;
 
    case WM_TIMER:
        InvalidateRect(hDlg, &invalidateRect, FALSE);
        return 0;
    }
 
    return FALSE;
}


Create a timer and define the necessary variables and handle window messages.



Run the program.


Check 'Canny Edge Detection'.


반응형
저작자표시 비영리 변경금지 (새창열림)

'Computer Vision' 카테고리의 다른 글

Detect, compute and draw keypoints - OpenCV 키포인트 찾기  (0) 2019.11.02
Screen capture - Windows API, OpenCV로 화면 캡쳐하기  (0) 2019.10.28
Convert Mat to Bitmap for Windows applications with GDI+ and surface crack detection - OpenCV Mat에서 비트맵으로 변환과 표면 크랙 검출  (0) 2019.10.26
Windows dialog based OpenCV Perspective Transform - 윈도우 다이얼로그 기반의 OpenCV Perspective Transform  (0) 2019.10.20
Simple motion detection with OpenCV - 간단한 움직임 감지  (2) 2019.09.15
Posted by J-sean
:

Convert Mat to Bitmap for Windows applications with GDI+ and surface crack detection - OpenCV Mat에서 비트맵으로 변환과 표면 크랙 검출

Computer Vision 2019. 10. 26. 10:06 |
반응형

GDI+ Bitmap class inherits from the Image class. The Image class provides methods for loading and saving vector images (metafiles) and raster images (bitmaps). You can build Windows applications with GDI+ and OpenCV.

 

 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
#include <windows.h>
#include <gdiplus.h>
#include <opencv2/opencv.hpp>
 
#pragma comment(lib, "gdiplus")
 
using namespace cv;
using namespace Gdiplus;
 
LRESULT CALLBACK WndProc(HWND, UINT, WPARAM, LPARAM);
HINSTANCE g_hInst;
LPCSTR lpszClass = "MatToBitmap";
 
Mat src;    // source image
Mat org;    // original image
Mat dst;    // processed image
  

Prepare necessities.

 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, LPSTR lpszCmdParam, int nCmdShow)
{
    HWND hWnd;
    MSG Message;
    WNDCLASS WndClass;
    g_hInst = hInstance;
 
    ULONG_PTR gpToken;
    GdiplusStartupInput gpsi;
    if (GdiplusStartup(&gpToken, &gpsi, NULL) != Ok) {
        MessageBox(NULL, TEXT("GDI+ start-up error."), TEXT("GDI+ Error"), MB_OK);
 
        return 0;
    }
 
    WndClass.cbClsExtra = 0;
    WndClass.cbWndExtra = 0;
    WndClass.hbrBackground = (HBRUSH)(COLOR_BTNFACE + 1); //(HBRUSH)GetStockObject(WHITE_BRUSH);
    WndClass.hCursor = LoadCursor(NULL, IDC_ARROW);
    WndClass.hIcon = LoadIcon(NULL, IDI_APPLICATION);
    WndClass.hInstance = hInstance;
    WndClass.lpfnWndProc = (WNDPROC)WndProc;
    WndClass.lpszClassName = lpszClass;
    WndClass.lpszMenuName = NULL;
    WndClass.style = CS_HREDRAW | CS_VREDRAW;
    RegisterClass(&WndClass);
 
    hWnd = CreateWindow(lpszClass, lpszClass, WS_OVERLAPPEDWINDOW,CW_USEDEFAULT, CW_USEDEFAULT,
        670, 340, NULL, (HMENU)NULL, hInstance, NULL);
    ShowWindow(hWnd, nCmdShow);
 
    while (GetMessage(&Message, 0, 0, 0)) {
        TranslateMessage(&Message);
        DispatchMessage(&Message);
    }
 
    GdiplusShutdown(gpToken);
 
    return (int)Message.wParam;
}
 
  

Start GDI+ and change the background color to the radio button color and Window size.

 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
bool getFile(PTCHAR filename)
{
    OPENFILENAME ofn;
    memset(&ofn, 0, sizeof(OPENFILENAME));
    ofn.lStructSize = sizeof(OPENFILENAME);
    ofn.hwndOwner = NULL;
    ofn.lpstrFilter = TEXT("all(*.*)\0*.*\0jpg(*.jpg)\0*.jpg\0png(*.png)\0*.png\0bmp(*.bmp)\0*.bmp\0");
    ofn.lpstrFile = filename;
    ofn.nMaxFile = MAX_PATH;
 
    if (GetOpenFileName(&ofn) != 0) {
        //MessageBox(NULL, filename, TEXT("File opened."), MB_OK);
 
        return true;
    }
    else {
        MessageBox(NULL, TEXT("File open failed"), TEXT("No file selected"), MB_OK);
 
        return false;
    }
}
 
  

Retrieve an image file name.

 

 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
bool ImageLoad()
{
    TCHAR filename[MAX_PATH] = "";
 
    if (getFile(filename)) {
        src = imread(filename);
        if (src.empty()) {
            MessageBox(NULL, TEXT("Image load failed"), TEXT("No image loaded"), MB_OK);
 
            return false;
        }
 
        return true;
    }
    else {
        return false;
    }
}
 
  

Read the image from the file.

 

1
2
3
4
5
6
7
8
void ImageResize()
{
    // maximum image display size: 320 X 240 with original ratio
    double ratio = min((double)320 / (double)src.size().width, (double)240 / (double)src.size().height);
    resize(src, src, cv::Size(), ratio, ratio);
    org = src.clone();
    cvtColor(org, org, COLOR_BGR2BGRA);
}
 
  

Resize an image.

 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
void ImageProcess(int direction)
{
    cvtColor(src, dst, COLOR_BGR2GRAY);
 
    float vertical[] = {
        -2, 0, 2,
        -2, 0, 2,
        -2, 0, 2
    };
 
    float horizontal[] = {
        -2, -2, -2,
        0, 0, 0,
        2, 2, 2
    };
 
    float diagonal[] = {
        -2, -2, 0,
        -2, 0, 2,
        0, 2, 2
    };
    
    Mat emboss;
    switch (direction)
    {
    case 0:
        emboss = Mat(3, 3, CV_32FC1, vertical);
 
        break;
 
    case 1:
        emboss = Mat(3, 3, CV_32FC1, horizontal);
 
        break;
 
    case 2:
        emboss = Mat(3, 3, CV_32FC1, diagonal);
 
        break;
 
    default:
        emboss = Mat(3, 3, CV_32FC1, vertical);
    }
    filter2D(dst, dst, -1, emboss, cv::Point(-1, -1), 128);
    
    cvtColor(dst, dst, COLOR_GRAY2BGRA);
}
 
  

Process an image with embossing filter.

 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
void OnPaint(HDC hdc)
{
    if (!dst.empty()) {
        Graphics G(hdc);
 
        Bitmap bitmapDst(dst.size().width, dst.size().height, dst.step, PixelFormat32bppARGB, dst.data);
        Bitmap bitmapOrg(org.size().width, org.size().height, org.step, PixelFormat32bppARGB, org.data);
        // stride(src.step): Integer that specifies the byte offset between the beginning of one scan line and
        // the next. This is usually(but not necessarily) the number of bytes in the pixel format(for example,
        // 2 for 16 bits per pixel) multiplied by the width of the bitmap. The value passed to this parameter
        // must be a multiple of four.
 
        G.DrawImage(&bitmapDst, 0, 0, bitmapDst.GetWidth(), bitmapDst.GetHeight());
        G.DrawImage(&bitmapOrg, bitmapOrg.GetWidth() + 10, 0, bitmapOrg.GetWidth(), bitmapOrg.GetHeight());
    }
    else {
        TextOut(hdc, 270, 110, TEXT("No image to display"), 19);
    }
}
 
  

Draw original and processed images.

 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
enum { ID_R1 = 101, ID_R2, ID_R3 };
HWND r1, r2, r3;
int x = 10;
int y = 270;
int w = 90;
int h = 20;
 
LRESULT CALLBACK WndProc(HWND hWnd, UINT iMessage, WPARAM wParam, LPARAM lParam)
{
    HDC hdc;
    PAINTSTRUCT ps;    
 
    switch (iMessage) {
    case WM_CREATE:
        if (ImageLoad()) {
            ImageResize();
            ImageProcess(0);
        }
        
        CreateWindow(TEXT("button"), TEXT("Filter type"), WS_CHILD | WS_VISIBLE |
            BS_GROUPBOX, 5, 250, 290, 50, hWnd, (HMENU)0, g_hInst, NULL);
 
        r1 = CreateWindow(TEXT("button"), TEXT("Vertical"), WS_CHILD | WS_VISIBLE |
            BS_AUTORADIOBUTTON | WS_GROUP, x, y, w, h, hWnd, (HMENU)ID_R1, g_hInst, NULL);
 
        r2 = CreateWindow(TEXT("button"), TEXT("Horizontal"), WS_CHILD | WS_VISIBLE |
            BS_AUTORADIOBUTTON, x + 90, y, w, h, hWnd, (HMENU)ID_R2, g_hInst, NULL);
 
        r3 = CreateWindow(TEXT("button"), TEXT("Diagonal"), WS_CHILD | WS_VISIBLE |
            BS_AUTORADIOBUTTON, x + 180, y, w, h, hWnd, (HMENU)ID_R3, g_hInst, NULL);
 
        CheckRadioButton(hWnd, ID_R1, ID_R3, ID_R1);
 
        return 0;
 
    case WM_COMMAND:
        if (!dst.empty()) {
            switch (LOWORD(wParam)) {
            case ID_R1:
                ImageProcess(0);
                break;
 
            case ID_R2:
                ImageProcess(1);
                break;
 
            case ID_R3:
                ImageProcess(2);
                break;
            }
 
            InvalidateRect(hWnd, NULL, TRUE);
        }
 
        return 0;
 
    case WM_PAINT:
        hdc = BeginPaint(hWnd, &ps);
        OnPaint(hdc);
        EndPaint(hWnd, &ps);
 
        return 0;
 
    case WM_DESTROY:
        PostQuitMessage(0);
 
        return 0;
    }
 
    return(DefWindowProc(hWnd, iMessage, wParam, lParam));
}
 
  

Define the necessary variables and handle window messages.

 

 

Run the application and select an image file.

 

Processed and original image with a vertical filter.

 

Processed and original image with a horizontal filter.

 

Processed and original image with a diagonal filter.

 

 

반응형
저작자표시 비영리 변경금지 (새창열림)

'Computer Vision' 카테고리의 다른 글

Screen capture - Windows API, OpenCV로 화면 캡쳐하기  (0) 2019.10.28
Windows dialog-based OpenCV program 윈도우 대화상자 OpenCV 프로그램 만들기  (2) 2019.10.26
Windows dialog based OpenCV Perspective Transform - 윈도우 다이얼로그 기반의 OpenCV Perspective Transform  (0) 2019.10.20
Simple motion detection with OpenCV - 간단한 움직임 감지  (2) 2019.09.15
3 ways to iterate through the matrix and comparison - OpenCV 행렬 원소를 모두 참조하는 세 가지 방법  (0) 2019.08.18
Posted by J-sean
:

Windows dialog based OpenCV Perspective Transform - 윈도우 다이얼로그 기반의 OpenCV Perspective Transform

Computer Vision 2019. 10. 20. 14:10 |
반응형

Build a windows dialog-based OpenCV perspective transform program.


Change Subsystem setting from Console to Windows if necessary.


Add a resource and create a dialog box.


Include and declare necessities.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
#include <Windows.h>
#include <opencv2/opencv.hpp>
#include "resource.h"
 
using namespace cv;
 
bool getFile(PTCHAR filename);
void onMouse(int event, int x, int y, int flags, void* userdata);
INT_PTR MainDlgProc(HWND hDlg, UINT iMessage, WPARAM wParam, LPARAM lParam);
 
HWND hDlgMain;
 
Mat sourceImage;
Mat targetImage;
// TTC card ratio
int width = 850;
int height = 550;
Point2f srcQuad[4], dstQuad[4];



Define WinMain which calls getFile() and DialogBox().

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
int APIENTRY WinMain(HINSTANCE hInstance, HINSTANCE hPrevInstance, LPSTR lpszCmdParam, int nCmdShow)
{
    TCHAR filename[MAX_PATH] = "";
    if (!getFile(filename)) {
        return -1;
    }
 
    sourceImage = imread(filename);
    if (sourceImage.empty()) {
        MessageBox(NULL, TEXT("Image load failed"), TEXT("No image loaded"), MB_OK);
 
        return -1;
    }
 
    namedWindow("sourceImage");
    setMouseCallback("sourceImage", onMouse);
 
    imshow("sourceImage", sourceImage);
 
    DialogBox(hInstance, MAKEINTRESOURCE(IDD_DIALOG1), HWND_DESKTOP, MainDlgProc);
 
    return 0;
}



getFile() retrieves an image file.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
bool getFile(PTCHAR filename)
{
    OPENFILENAME ofn;
    memset(&ofn, 0, sizeof(OPENFILENAME));
    ofn.lStructSize = sizeof(OPENFILENAME);
    ofn.hwndOwner = NULL;
    ofn.lpstrFilter = TEXT("all(*.*)\0*.*\0jpg(*.jpg)\0*.jpg\0png(*.png)\0*.png\0bmp(*.bmp)\0*.bmp\0");
    ofn.lpstrFile = filename;
    ofn.nMaxFile = MAX_PATH;
 
    if (GetOpenFileName(&ofn) != 0) {
        //MessageBox(NULL, filename, TEXT("File opened."), MB_OK);
 
        return true;
    }
    else {
        MessageBox(NULL, TEXT("File open failed"), TEXT("No file selected"), MB_OK);
 
        return false;
    }
}



Define a mouse callback function that processes perspective transform.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
void onMouse(int event, int x, int y, int flags, void* userdata)
{
    static int cnt = 0;
 
    if (event == EVENT_LBUTTONDOWN) {
        if (cnt < 4) {
            srcQuad[cnt++] = Point2f(x, y);
 
            circle(sourceImage, Point2f(x, y), 3, Scalar(0, 0, 255), -1);
            imshow("sourceImage", sourceImage);
 
            if (cnt == 4) {
                /*
                width = max(
                    sqrt(pow(srcQuad[0].x - srcQuad[1].x, 2)) + sqrt(pow(srcQuad[0].y - srcQuad[1].y, 2)),
                    sqrt(pow(srcQuad[2].x - srcQuad[3].x, 2)) + sqrt(pow(srcQuad[2].y - srcQuad[3].y, 2))
                );
                height = max(
                    sqrt(pow(srcQuad[1].x - srcQuad[2].x, 2)) + sqrt(pow(srcQuad[1].y - srcQuad[2].y, 2)),
                    sqrt(pow(srcQuad[3].x - srcQuad[0].x, 2)) + sqrt(pow(srcQuad[3].y - srcQuad[0].y, 2))
                );
                */            
 
                dstQuad[0] = Point2f(0, 0);
                dstQuad[1] = Point2f(width - 1, 0);
                dstQuad[2] = Point2f(width - 1, height - 1);
                dstQuad[3] = Point2f(0, height - 1);
 
                Mat perspectiveTransform = getPerspectiveTransform(srcQuad, dstQuad);
 
                warpPerspective(sourceImage, targetImage, perspectiveTransform, Size(width, height));
 
                cnt = 0;
                
                imshow("targetImage", targetImage);
            }
        }
    }
}



Dialog box procedure sets and/or retrieves target image width and height value from the dialog. 

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
INT_PTR MainDlgProc(HWND hDlg, UINT iMessage, WPARAM wParam, LPARAM lParam)
{
    switch (iMessage) {
    case WM_INITDIALOG:
        hDlgMain = hDlg;
        SetDlgItemInt(hDlg, IDC_WIDTH, width, FALSE);
        SetDlgItemInt(hDlg, IDC_HEIGHT, height, FALSE);
 
        return TRUE;
 
    case WM_COMMAND:
        switch (LOWORD(wParam)) {
        case IDOK:
            width = GetDlgItemInt(hDlg, IDC_WIDTH, NULL, FALSE);
            height = GetDlgItemInt(hDlg, IDC_HEIGHT, NULL, FALSE);
 
            return TRUE;
 
        case IDCLOSE:
            EndDialog(hDlg, IDOK);
 
            return TRUE;
        }
    }
 
    return FALSE;
}




Run the program and select an image file.


You can change the target image size or just leave it.


Source image window.


Choose four corners from the left-top to clockwise.


Perspective transform processed image.


반응형
저작자표시 비영리 변경금지 (새창열림)

'Computer Vision' 카테고리의 다른 글

Windows dialog-based OpenCV program 윈도우 대화상자 OpenCV 프로그램 만들기  (2) 2019.10.26
Convert Mat to Bitmap for Windows applications with GDI+ and surface crack detection - OpenCV Mat에서 비트맵으로 변환과 표면 크랙 검출  (0) 2019.10.26
Simple motion detection with OpenCV - 간단한 움직임 감지  (2) 2019.09.15
3 ways to iterate through the matrix and comparison - OpenCV 행렬 원소를 모두 참조하는 세 가지 방법  (0) 2019.08.18
How to make and use a roation matrix and a translation matrix - OpenCV 회전 행렬, 이동 행렬 만들기  (0) 2019.08.13
Posted by J-sean
:

Simple motion detection with OpenCV - 간단한 움직임 감지

Computer Vision 2019. 9. 15. 17:18 |
반응형

Detects motions in the vision and saves an image of it as a PNG file. Adjust the sensitivity value if needed.


카메라 영상의 움직임을 감시하고 움직임이 감지된 순간의 영상을 저장한다. sensitivity 값으로 감도를 조절한다.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
#include <opencv2/opencv.hpp>
 
using namespace std;
using namespace cv;
 
int main(int argc, char** argv)
{
    VideoCapture cap(0);
 
    Mat frameNew;
    Mat frameOld;
    Mat frameDiff;
 
    double min, max;
    int sensitivity = 100;
    int detectionCount = 0;
 
    cap >> frameOld;
 
    while (true)
    {
        cap >> frameNew;
        if (frameNew.empty())
            break;
 
        // Calculates the per-element absolute difference
        // between two arrays or between an array and a scalar.
        absdiff(frameNew, frameOld, frameDiff);
        cvtColor(frameDiff, frameDiff, COLOR_BGR2GRAY);
        minMaxLoc(frameDiff, &min, &max);
 
        if (max > sensitivity)
        {
            cout << "Motion detected. (Max: " << max << ")" << endl;
 
            // For PNG, it can be the compression level from 0 to 9.
            // A higher value means a smaller size and longer compression time.
            vector<int> compression_params;
            compression_params.push_back(IMWRITE_PNG_COMPRESSION);
            compression_params.push_back(3);
            if (imwrite(format("detection_%03d.png", detectionCount++), frameNew, compression_params))
                cout << "Image saved." << endl;
            else
                cout << "Image not saved." << endl;
        }
 
        imshow("Motion Detectoion", frameDiff);
        
        frameNew.copyTo(frameOld);
 
        if (waitKey(10) == 27)
            break;
    }
 
    return 0;
}
cs



<frameDiff>


<Detection_XXX.png>




반응형
저작자표시 비영리 변경금지 (새창열림)

'Computer Vision' 카테고리의 다른 글

Convert Mat to Bitmap for Windows applications with GDI+ and surface crack detection - OpenCV Mat에서 비트맵으로 변환과 표면 크랙 검출  (0) 2019.10.26
Windows dialog based OpenCV Perspective Transform - 윈도우 다이얼로그 기반의 OpenCV Perspective Transform  (0) 2019.10.20
3 ways to iterate through the matrix and comparison - OpenCV 행렬 원소를 모두 참조하는 세 가지 방법  (0) 2019.08.18
How to make and use a roation matrix and a translation matrix - OpenCV 회전 행렬, 이동 행렬 만들기  (0) 2019.08.13
Measuring execution time of a function - OpenCV에서 실행 시간 측정 하기  (0) 2019.07.14
Posted by J-sean
:

3 ways to iterate through the matrix and comparison - OpenCV 행렬 원소를 모두 참조하는 세 가지 방법

Computer Vision 2019. 8. 18. 10:29 |
반응형

There are 3 ways to iterate through the matrix. Below code shows how to use them and compares their speed.

행렬 원소를 모두 참조하는 세 가지 방법을 보여주고 속도를 비교 합니다.


1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
#include <opencv2/opencv.hpp>
 
using namespace std;
using namespace cv;
 
int main(int argc, char** argv)
{
    TickMeter tm;
    Mat mat = Mat::zeros(600, 800, CV_8UC1);
 
    tm.start();
    for (int i = 0; i < mat.rows; i++)
        for (int j = 0; j < mat.cols; j++)
            mat.at<uchar>(i, j)++;
    tm.stop();
    cout << ".at<> execution time: " << tm.getTimeSec() << endl;
    tm.reset();    
 
    tm.start();
    for (int i = 0; i < mat.rows; i++)
    {
        uchar* p = mat.ptr<uchar>(i);
        for (int j = 0; j < mat.cols; j++)
            p[j]++;
    }
    tm.stop();
    cout << ".ptr<> execution time: " << tm.getTimeSec() << endl;
    tm.reset();
 
    tm.start();
    for (MatIterator_<uchar> it = mat.begin<uchar>(); it != mat.end<uchar>(); it++)
        (*it)++;
    tm.stop();
    cout << "MatIterator execution time: " << tm.getTimeSec() << endl;
    tm.reset();
 
    return 0;
}
cs





반응형
저작자표시 비영리 변경금지 (새창열림)

'Computer Vision' 카테고리의 다른 글

Windows dialog based OpenCV Perspective Transform - 윈도우 다이얼로그 기반의 OpenCV Perspective Transform  (0) 2019.10.20
Simple motion detection with OpenCV - 간단한 움직임 감지  (2) 2019.09.15
How to make and use a roation matrix and a translation matrix - OpenCV 회전 행렬, 이동 행렬 만들기  (0) 2019.08.13
Measuring execution time of a function - OpenCV에서 실행 시간 측정 하기  (0) 2019.07.14
Random Number Generator(RNG) - OpenCV 난수 발생기 사용  (0) 2019.07.13
Posted by J-sean
:

티스토리툴바