[RELEASED] OpenCV for Unity

Hello everyone!

Does anyone know how I can add faces from image files and then do the recognition from the webcam?

 

Hi there -- all the examples of object detection I'm finding are about faces. However, we need something that would find other random objects in the environment from an AR camera feed. Is it possible to use the OpenCV training functions to do our own object detection training with this plugin? like opencv_traincascade to build the xml file from a bunch of images, then custom_cascade.DetectMultiScale()? Are these or something like them exposed with OpenCVForUnity? (the api showed me some methods with "train" in the name, but I didn't see those)

 

Have you already tried the RealTime FaceRecognition Example?
https://assetstore.unity.com/packages/templates/tutorials/realtime-facerecognition-example-73233

 

OpenCVForUnity itself does not have the ability to train cascade files.
Of course, it is possible to use OpenCVForUnity for object detection with cascade files trained using opencv_traincascade.
https://github.com/opencv/opencv/bl...c3/doc/tutorials/others/traincascade.markdown

 

Hello,
Your asset seems very intersting, and I have a question : how complicated would it be to achieve this :

? Maybe someone already made something like that ?
Best regards, Phil

 

Ah okay, so I would need to write my own plugin to have access to the traincascade function?

 

YoloObjectDetectionWebCamTextureExample is included in OpenCVForUnity. This is an example of detecting a car from a WebCam feed.

 

That’s correct.

 

I am trying to convert a RenderTexture directly to a Mat using a Ptr or NativeArray, but haven't had any luck. I was hoping you might be able to point me in the right direction.

The pasted code is a sample of some of my attempts.

The second bit is for a Texture2D which I was using to try and simplify the problem. It works, but not sure how to get a similar native array for render textures.

Code (CSharp):

1.  
2. private RenderTexture _myRenderTexture;
3. private Texture2D _myTexture2D;
4. public void PtrToMat()
5.  {
6.       var targetMat = new Mat(1920, 1080, CvType.CV_8UC4);
7.        
8.       //This doesn't work and RenderTextures don't seem to support GetRawTextureData
9.       var intPtr = _myRenderTexture.GetNativeTexturePtr();
10.       MatUtils.copyToMat(intPtr,targetMat);
11.        
12.       //This works for a Texture2D
13.       var nativeArray = _myTexture2D.GetRawTextureData<Color32>();
14.       MatUtils.copyToMat(nativeArray,targetMat);
15.  
16.  }

Thanks in advance.

Also, I just tried the NativeArrawUnsafeUtility and probably don't know what I'm doing...Which seems safe.

Code (CSharp):

1. var ptr = _myRenderTexture.GetNativeTexturePtr();
2.  
3. NativeArray<Color32> rtNative = NativeArrayUnsafeUtility.ConvertExistingDataToNativeArray<Color32>((void*) ptr, 1920 * 1080 * 4,Allocator.Persistent);
4.  
5. MatUtils.copyToMat(rtNative,targetMat);
6.  

The rtNative array never seems to get populated.

 

The ponters obtained from the GetNativeTexturePtr() method cannot be used for MatUtils.copyToMat(intPtr, mat) since they do not indicate the first pointer of the pixel buffer in memory.
We looked into the latest information on this subject some time ago, but unfortunately we have not found a more efficient way to get pixel data from a RenderTexture other than "ReadPixels" or "AsyncGPUReadback".

 

Need Help !!!

Unity Editor crash because of opencvforunity.dll many times.

 

Thank you!

 

Thank you very much for reporting.

Could you tell me the environment you tested?
OpenCV for Unity version :
Unity version :
Windows OS version :

 

Unity version: 2020.3.11f1 and 2021.1.9f1
OpenCV for Unity version: v2.4.4
Windows OS version: Windows 10 20H2 19042.985

 

In my environment, OpenCVForUnity is running on multiple PCs without any problems.
Unity version: 2020.3.11f1
OpenCV for Unity version: v2.4.4
Windows OS version: Windows 10 20H2 19042.1052

Also make sure that you have enough free disk space.

 

Can this asset be used with remote cameras streaming over either UDP or TCP, or would this require some 3rd party program to decode the stream and re-broadcast it as a faked webcam device?

 

To play streaming file, ffmpeg.dll is required.
1)Download "OpenCV for Windows Version 4.5.2"(https://sourceforge.net/projects/opencvlibrary/).
2)Copy "opencv_videoio_ffmpeg452_64.dll" to the root directory of the Project.

3)Edit User Variables.
https://stackoverflow.com/questions...reply-when-cv2-videocapture-rtsp-onvif-camera


I succeeded in playing this file.

Code (CSharp):

1.             //capture.open(Utils.getFilePath(VIDEO_FILENAME));
2.             capture.open("rtsp://wowzaec2demo.streamlock.net/vod/mp4:BigBuckBunny_115k.mov", Videoio.CAP_FFMPEG);

 

Hey @EnoxSoftware ,
I recently purchased your asset. Im really liking it so far. I am using the resnet face detetction example and trying to integrate a customizable-scifi-holo-interface (link:https://assetstore.unity.com/packages/2d/textures-materials/customizable-scifi-holo-interface-69794) to it instead of the green bounding boxes. (Pls see attached pictures).

According to a video I was watching, I was to modify the code of "DnnObjectDetectionWebCamTextureExample.cs" by brining in this GUI object. Here is the github link for the modified code as provided by the video(https://bit.ly/2TksAlw). However I got some errors after changing the code.
Assets/OpenCVForUnity/Examples/MainModules/dnn/LibFaceDetectionV3Example/LibFaceDetectionV3WebCamTextureExample.cs(71,33): error CS0115: 'LibFaceDetectionV3WebCamTextureExample.postprocess(Mat, List<Mat>, Net, int)': no suitable method found to override.

youtube video link:

.

Thanks so much in advance.

Edit: I am using unity 2020.3.13f1.
OpenCVforUnity 2.4.2

 

Hello EnoxSoftware,

I playing with the YoloObjectDetectionWebCamTextureExample its work nice, but I would like to detect only "Person" and in just specific area(rectangle area) of the video input. I try something like this in the postprocess() first for just show the "Person" green rectangles I trying this

Code (CSharp):

1.  
2. if(classIdsList[idx].ToString() == "0"){
3.                     //Debug.Log("personas: " + idx.ToString());
4.                     Rect2d box = boxesList[idx];
5.                     drawPred(classIdsList[idx], confidencesList[idx], box.x, box.y,
6.                     box.x + box.width, box.y + box.height, frame);
7.                 }
8.  

It works but I sure is not the correct way, after I would like to detect if is no "Person" detected for show some message, I try this

Code (CSharp):

1.  
2. OpenCVForUnity.CoreModule.Rect2d[] rects = boxesList.ToArray();
3.                     for (int i = 0; i < rects.Length; i++)
4.                     {
5.                         Debug.Log (rects);
6.                         if ( rects.Length <= 0 )
7.                         {
8.                             Debug.Log("No person detected");
9.                         }
10.                     }
11.  

but no luck, can you help me? how can trigger a flag if is not person detected.
Also if is not too much to ask, How can limited the area of recognized in order to no "scan" all the frame just a certain area of the background video

Thx for awesome work,

 

The video seems to be based on the example code included in an older version of OpenCVForUnity.
I have created a modified version of the latest example code as per the instructions in the video and attached it so you can try it out.

 

Try replacing the code near line 606 of the DocumentDetectionWebCamTextureExample.cs with this code.

Code (CSharp):

1.  
2.             bool personDetected = false;
3.             for (int idx = 0; idx < boxesList.Count; ++idx)
4.             {
5.                 if (classIdsList[idx] == 0)
6.                 {
7.                     //Debug.Log("personas: " + idx.ToString());
8.                     Rect2d box = boxesList[idx];
9.                     drawPred(classIdsList[idx], confidencesList[idx], box.x, box.y,
10.                         box.x + box.width, box.y + box.height, frame);
11.                     personDetected = true;
12.                 }
13.             }
14.             if (!personDetected)
15.             {
16.                 Debug.Log("No person detected");
17.             }
18.  

 

Hi @EnoxSoftware,
Thanks so much for the reply. I tried it however it doesnt seem to be doing much. It doesnt seem to detect any face. It just shows like this(With the cube encircled). Im very new to c# thats why its really hard for me.

The main idea was to use the game object as a symbol of a face detection(job of the green bounding boxes). So I want to be able to replace the green bounding boxes with the game object as soon as it detects a face. Hope you understand where I am going.

Thanks for all the help

 

According to the message in the screenshot, it seems that the model file YuFaceDetectNet.onnx does not exist in the project.
In order to run the LibFaceDetectionV3Example, you will need to download the model files from this link and place them in the "Assets/StreamingAsset/dnn/" folder.
https://github.com/ShiqiYu/libfacedetection.train/raw/master/tasks/task1/onnx/YuFaceDetectNet.onnx

 

I installed it and moved it to "Assets/StreamingAsset/dnn/" folder, but it results into Unity Crashing.

Thanks once again.

 

Hi @EnoxSoftware,
I`m trying to run Aruco.detectMarkers () with AprilTag's corner refinement but it throws me an "Integer division by zero" exception in UWP on Hololens2. I am on version 2.4.4 of OpenCv for Unity and OpenCV4.5.2.

As I've been reading https://github.com/opencv/opencv_contrib/issues/2643, it may be that an older version of OpenCV works, and I was wondering if you could send me an older version that has an OpenCV version lower than 4.3.0, so I can test it. Otherwise, could you tell me how to run Aruco.detectMarkers () with AprilTag's corner refinement in Hololens2?

 

Thank you very much for reporting.
Could you send me an email using the contact form below?
https://enoxsoftware.com/opencvforunity/contact/technical-inquiry/

 

@EnoxSoftware pls help

 

I just tested it in a similar environment to yours and it seems to be working fine. (I am using unity 2020.3.13f1 and OpenCVforUnity 2.4.4)
Could you please let me know if the other examples work after completing the setup according to the "ReadMe.pdf" and "StreamingAssets/dnn/setup_dnn_module.pdf" included with OpenCVForUnity?

 

Hey @EnoxSoftware,
I tested each and every example in the "StreamingAssets/dnn/setup_dnn_module.pdf". Everything worked except for 3.

• MaskRCNNExample(error message in the picture)
• LibFaceDetectionV3Example(unity just crashes)
• LibFaceDetectionV2Example(links dont work)

I have attached pictures pls check.

 

Due to a mistake on my part, I have given the wrong answer to that.
It seems that the content of the distribution site for the LibFaceDetectionV2Example and LibFaceDetectionV3Example model files has recently changed.
The correct URL is here.

LibFaceDetectionV2Example:
https://github.com/ShiqiYu/libfaced...dels/caffe/yufacedetectnet-open-v2.caffemodel
https://github.com/ShiqiYu/libfaced...models/caffe/yufacedetectnet-open-v2.prototxt

LibFaceDetectionV3Example:
https://github.com/ShiqiYu/libfaced...c801dd6/tasks/task1/onnx/YuFaceDetectNet.onnx

As for MaskRCNNExample, I checked it again, but there was no problem.

I apologize for the extra trouble this has caused.

 

@EnoxSoftware,

It finally works thanks to your Help. Thanks so much for this. It seems to be that link that caused the trouble. Don't worry about the trouble, It helped me try out the other examples.

Thanks so much once again.

 

Hey, @EnoxSoftware, I am trying to input the Human Segmentation Texture2D from the Occlusion Manager of ARFoundation to FloodFill with no success. the Issue, I believe is on the conversion from texture2D to Mat. the code I am using is the following:

Code (CSharp):

1.         void ProcessWithOpenCV(Texture2D texture)
2.         {
3.             Texture2D p_texture = texture;
4.          
5.             // Mat mat = new Mat (p_texture.width, p_texture.height, CvType.CV_8UC3);
6.             Mat mat = new Mat ();
7.  
8.             Utils.fastTexture2DToMat(p_texture, mat, true);
9.  
10.             Imgproc.cvtColor(mat,mat,Imgproc.COLOR_RGB2GRAY);
11.  
12.             ProcessImageFloodFill(mat);
13.  
14.             Utils.fastMatToTexture2D(mat,p_texture,true,1,true,true,true);
15.  
16.             m_Texture = p_texture;
17.          
18.             ocv_rawImage.texture = m_Texture;
19.  
20.         }
21.  
22. void ProcessImageFloodFill(Mat mat)
23. {
24. //Clona matrix para uma nova para efeito de processamento
25. Mat im_Threshold = mat.clone();
26. //Aplica um gaussiam blur e em seguida um Trheshold
27. Imgproc.GaussianBlur(im_Threshold,im_Threshold,new Size(5,5),0);
28. Imgproc.threshold(im_Threshold,im_Threshold,220,255,Imgproc.THRESH_OTSU);
29. //Clona matrix para uma nova para efeito de processamento
30. Mat im_threshold_inv = im_Threshold.clone();
31. //Inverte o Threshold e copia para a Matrix original
32. Core.bitwise_not(im_Threshold,im_threshold_inv);
33.  
34. Mat im_floodFill = im_threshold_inv.clone();
35. Imgproc.GaussianBlur(im_floodFill,im_floodFill,new Size(5,5),0);
36. Imgproc.floodFill(im_floodFill,im_floodFill, new Point(0,0), new Scalar(0));
37. im_floodFill.copyTo(mat);
38. Imgproc.cvtColor(mat,mat,Imgproc.COLOR_GRAY2RGB);
39. }
40.  
41.  

the commented line of creating a Mat was crashing the app. Any ideas of what am I doing wrong?

 

I found a code sample that converts ARFoundation's XRCameraImage to OpenCV's Mat here.
https://gitlab.com/-/snippets/2041871
I haven't tried it in detail, but it may be helpful.

 

I'll give it a try. Although my problem is not with the camera image but with the Human segmentation but it may give me some clues. Thanks!

 

Hi, I was wondering if there is a tutorial or a document where it could explain me how to use custom descriptors in opencv unity, i was reading the documentation and you have this metod load

Code (CSharp):

1. string svmFile = Utils.getFilePath("svm.xml");
2. Debug.Log(svmFile);
3. des.load(svmFile);
4.  
5. //tested too but does not work (unity freezes)
6. string svmFile = Utils.getFilePath("svm.xml");
7. bytes = System.IO.File.ReadAllBytes(svmFile);
8. MatOfByte mat = new MatOfByte(bytes);
9. des.setSVMDetector(MatOfByte.fromNativeAddr(mat.nativeObj));
10. // or this one
11. des.setSVMDetector(mat);

but for me it does not work, already tested with other tipes of svm files and none worked, someone can help me? I want to use the hog detector with the camera too but that was easy, I only need the way to use custom descriptors in opencv unity.

 

I added a short piece of code to the example "Assets\OpenCVForUnity\Examples\MainModules\ml\SVMExample\SVMExample.cs" to test saving and loading the SVM training data file.
It appears to be working properly.

Code (CSharp):

1. using UnityEngine;
2. using UnityEngine.SceneManagement;
3. using System.Collections;
4. using OpenCVForUnity.CoreModule;
5. using OpenCVForUnity.MlModule;
6. using OpenCVForUnity.ImgprocModule;
7. using OpenCVForUnity.UnityUtils;
8.  
9. namespace OpenCVForUnityExample
10. {
11.     /// <summary>
12.     /// SVM Example
13.     /// An example of  find a separating straight line using the Support Vector Machines (SVM).
14.     /// Referring to http://docs.opencv.org/3.1.0/d1/d73/tutorial_introduction_to_svm.html#gsc.tab=0.
15.     /// </summary>
16.     public class SVMExample : MonoBehaviour
17.     {
18.         // Use this for initialization
19.         void Start ()
20.         {
21.             // Data for visual representation
22.             int width = 512, height = 512;
23.             Mat image = Mat.zeros (height, width, CvType.CV_8UC4);
24.  
25.             // Set up training data
26.             int[] labels = { 1, -1, -1, -1 };
27.             float[] trainingData = { 501, 10, 255, 10, 501, 255, 10, 501 };
28.             Mat trainingDataMat = new Mat (4, 2, CvType.CV_32FC1);
29.             trainingDataMat.put (0, 0, trainingData);
30.             Mat labelsMat = new Mat (4, 1, CvType.CV_32SC1);
31.             labelsMat.put (0, 0, labels);
32.  
33.             // Train the SVM
34.             SVM svm = SVM.create ();
35.             svm.setType (SVM.C_SVC);
36.             svm.setKernel (SVM.LINEAR);
37.             svm.setTermCriteria (new TermCriteria (TermCriteria.MAX_ITER, 100, 1e-6));
38.             svm.train (trainingDataMat, Ml.ROW_SAMPLE, labelsMat);
39.  
40.             // Show the decision regions given by the SVM
41.             byte[] green = { 0, 255, 0, 255 };
42.             byte[] blue = { 0, 0, 255, 255 };
43.             for (int i = 0; i < image.rows (); ++i)
44.                 for (int j = 0; j < image.cols (); ++j) {
45.                     Mat sampleMat = new Mat (1, 2, CvType.CV_32FC1);
46.                     sampleMat.put (0, 0, j, i);
47.            
48.                     float response = svm.predict (sampleMat);
49.                     if (response == 1)
50.                         image.put (i, j, green);
51.                     else if (response == -1)
52.                         image.put (i, j, blue);
53.                 }
54.  
55.             // Show the training data
56.             int thickness = -1;
57.             int lineType = 8;
58.        
59.             Imgproc.circle (image, new Point (501, 10), 5, new Scalar (0, 0, 0, 255), thickness, lineType, 0);
60.             Imgproc.circle (image, new Point (255, 10), 5, new Scalar (255, 255, 255, 255), thickness, lineType, 0);
61.             Imgproc.circle (image, new Point (501, 255), 5, new Scalar (255, 255, 255, 255), thickness, lineType, 0);
62.             Imgproc.circle (image, new Point (10, 501), 5, new Scalar (255, 255, 255, 255), thickness, lineType, 0);
63.  
64.             // Show support vectors
65.             thickness = 2;
66.             lineType = 8;
67.             Mat sv = svm.getUncompressedSupportVectors ();
68. //                      Debug.Log ("sv.ToString() " + sv.ToString ());
69. //                      Debug.Log ("sv.dump() " + sv.dump ());
70.             for (int i = 0; i < sv.rows (); ++i) {
71.                 Imgproc.circle (image, new Point ((int)sv.get (i, 0) [0], (int)sv.get (i, 1) [0]), 6, new Scalar (128, 128, 128, 255), thickness, lineType, 0);
72.             }
73.  
74.  
75.             Texture2D texture = new Texture2D (image.width (), image.height (), TextureFormat.RGBA32, false);
76.             Utils.matToTexture2D (image, texture);
77.             gameObject.GetComponent<Renderer> ().material.mainTexture = texture;
78.  
79.  
80.  
81.             //////////////////////////////////////
82.             // Save and Load SVM trained data file Test
83.  
84.             Utils.setDebugMode(true);
85.  
86.             string SAVE_TRAINED_DATA_PATH = Application.persistentDataPath + "/svm_trained_data.xml";
87.  
88.             svm.save(SAVE_TRAINED_DATA_PATH);
89.             //Debug.Log("file path: " + SAVE_TRAINED_DATA_PATH);
90.  
91.  
92.             SVM test_svm = SVM.load(SAVE_TRAINED_DATA_PATH);
93.             // SVM test_svm = SVM.load(Utils.getFilePath("_svm_test/svm_trained_data.xml"); // When loading from the StreamingAssets folder.
94.  
95.             Debug.Log("isTrained: " + test_svm.isTrained());
96.  
97.             if (test_svm.isTrained())
98.             {
99.                 Mat predicted = new Mat();
100.                 test_svm.predict(trainingDataMat, predicted);
101.  
102.                 Debug.Log("test labels: " + labelsMat.dump());
103.                 Debug.Log("predicted: " + predicted.dump());
104.             }
105.  
106.             Utils.setDebugMode(false);
107.  
108.             //
109.             // test labels: [1; -1; -1; -1]
110.             // predicted: [1; -1; -1; -1]
111.             //
112.  
113.             //////////////////////////////////////
114.         }
115.  
116.         // Update is called once per frame
117.         void Update ()
118.         {
119.  
120.         }
121.  
122.         /// <summary>
123.         /// Raises the back button click event.
124.         /// </summary>
125.         public void OnBackButtonClick ()
126.         {
127.             SceneManager.LoadScene ("OpenCVForUnityExample");
128.         }
129.     }
130. }

 

We are having the same issue... did you already figure out why is it happening ?

 

ok, maybe it's because I’m training in Dlib using python, or opencv in python, I’m not training in unity opencv because I don't really understand how I can train using images, I can see you are using just an array of values as training data, but I need an array of images, could you tell me how can I accomplice that?

 

I think Unity's opencv is compatible with python's opencv SVM training data.

Here's how to flatten an image into an array of values.

Code (CSharp):

1.  
2. Mat img = Imgcodecs.imread(Utils.getFilePath("svm/traning_image01.png"), Imgcodecs.IMREAD_GRAYSCALE);
3. Imgproc.resize(img, img, new Size(64,64));
4. Mat flatImg = img.reshape(1, 1); // to flatten
5. flatImg.convertTo(flatImg, CvType.CV_32FC1);
6.  

 

I am still investigating this issue.
Could you send me the code you've tested?
https://enoxsoftware.com/opencvforunity/contact/technical-inquiry/

 

Hi for my client I need to create a WebGL version with face recognition.
I downloaded and imported the package in an empty project version 2019.1.6f1
and built the example scene:

FaceDetectionWebCamTextureExample

But I got this error on the play


Any way to solve this?

 

Thank you very much for reporting.
Could you tell me the environment you tested?
OpenCV for Unity version :
Unity version :
browser :

 

yes, sure:

Unity version: 2019.1.6
Opencv Version: 2.4.2
Browser Google Chrome/Firefox

Please help me

 

Hi @EnoxSoftware

When running the WebCamTextureMarkerBasedARExample my AR Object flicks when keeping the marker still. See attached Animation2.gif for a screencast.

I wrote several variable-values to a csv file when debugging. I found that the Vector3 forward and upwards produces by ARUtils.ExtractRotationFromMatrix() seems to be changing rapidly and I assume this is what causes my object to make sudden rotations. Highlighted in red is what I am referring to.



Do you know what is causing this?

The problem only seems to occur when holding the marker directly facing the camera. When the marker is tilted a bit, the problem disappears.

A noise reduction algorithm did not work since the shifts are to big. Any tips of how to mitigate/prevent this is much appreciated since this is for a client.

 

It worked fine in my environment. I would like to send you the project I have set up, could you send me an email using the form below?
https://enoxsoftware.com/opencvforunity/contact/technical-inquiry/

 

Calling the calcOpticalFlowPyrLK function in OpenCV for Unity WebGL ~ 20-100 times slower than calling the same function in Android version or via the official version opencv.js. Similarly poor performance with other functions such as estimateAffinePartial2D.

Function calcOpticalFlowPyrLK execution time (tested on the same Android device):
OpenCV for Unity WebGL ~ 270 ms
OpenCV for Unity Android ~ 10 ms
opencv.js ~ 12 ms (can be run from Unity WebGL too)

In a real project, the time can increase up to 700 ms or more (when the camera is running and other functions are called). At the same time, performance in Android version or in opencv.js does not drop.

How to improve the performance of these functions or are they impossible to use in real time and should be used opencv.js?

Code (CSharp):

1.     // Init
2.     Mat localPrevMat, localNextMat;
3.     MatOfPoint2f mOP2fPrevTrackPts = new MatOfPoint2f();
4.     MatOfPoint2f mOP2fNextTrackPts = new MatOfPoint2f();
5.     MatOfByte status = new MatOfByte();
6.     MatOfFloat err = new MatOfFloat();
7.     private readonly Size winSize = new Size(22, 7);
8.     private readonly TermCriteria criteria = new TermCriteria(TermCriteria.EPS + TermCriteria.COUNT, 10, 0.03);
9.  
10.  
11.         localPrevMat = new Mat(160, 160, CvType.CV_8UC4);
12.         localNextMat = new Mat(160, 160, CvType.CV_8UC4);
13.  
14.  
15.  
16.         // Call in Update
17.         Point[] localPrevPoints = new Point[500];
18.         Point[] localNextPoints = new Point[500];
19.         for (int i = 0; i < localPrevPoints.Length; i++)
20.         {
21.             localPrevPoints[i] = new Point(UnityEngine.Random.Range(0, 159), UnityEngine.Random.Range(0, 159));
22.             localNextPoints[i] = new Point(localPrevPoints[i].x, localPrevPoints[i].y);
23.         }
24.  
25.         mOP2fPrevTrackPts.fromArray(localPrevPoints);
26.         mOP2fNextTrackPts.fromArray(localNextPoints);
27.  
28.         System.Diagnostics.Stopwatch sw = new System.Diagnostics.Stopwatch();
29.         sw.Restart();
30.         Video.calcOpticalFlowPyrLK(localPrevMat, localNextMat, mOP2fPrevTrackPts, mOP2fNextTrackPts, status, err, winSize, 5, criteria, Video.OPTFLOW_USE_INITIAL_FLOW, 1e-4);
31.         sw.Stop();
32.         Debug.Log(("calcOpticalFlowPyrLK: ", sw.ElapsedMilliseconds));

 

Hi, i sended you the email with that module.... let me know....

 

MarkerBasedARExample Code is a rewrite of https://github.com/MasteringOpenCV/code/tree/master/Chapter2_iPhoneAR/Example_MarkerBasedAR using “OpenCV for Unity”. The algorithm is described in detail in "Mastering OpenCV with Practical Computer Vision Projects". http://www.packtpub.com/cool-projects-with-opencv/book

This example is a tutorial code, so it is not very accurate. You may be able to get better accuracy by using the ArUco class instead.
https://github.com/EnoxSoftware/Ope...uco/ArUcoExample/ArUcoWebCamTextureExample.cs

 

Thank you very much for reporting.
Could you send me an email using the contact form below?
https://enoxsoftware.com/opencvforunity/contact/technical-inquiry/

 

Hi,
I sent you an email.