Fast Opencv people pedestrian detection Tutorial by CascadeClassifier
Simple Opencv C++ tutorial and example of people detection in video samples and pictures. People detection performance in opencv could be enhanced with simple steps. This text uses old detection methods by CascadeClassifier. The current software will definitely take advantage of DNN modules and deep neural networks that outperform the old approach by haar and LBP cascades. You can find this topic on my block here.
Opencv tutorial installation of OpenCV
Update
from 2024.
You can simply prepare the project inside the Visual Studio 2015 by Nuget Packages. This approach is easy for beginners and better than standard installation with all the environmental variables problems. Just follow the installation steps here
Opencv is a great and complex tool. There are a lot of image processing and also machine learning features. You can simply learn your own detector. I would like to prepare some tutorials on how to learn your own detector. It is a long time run.
All, you need to do, is have some experience and basic opencv tools.
under opencv/build/x64/vc14/bin
opencv_createsamples.exe
opencv_traincascade.exe
Prepare your dataset and files with labels that are better under Linux. You can use simple shell scripts.
This is really for long tutorial, but you can do image annotation in Windows as well and maybe use the new tool opencv_annotation,exe, but i don't have any experience with this. My own scripts and programs crop, save and annotate data. I have my own image cropper to prepare datasets. You can build your own dataset in a few days. It is tricky too.
First step
Use basic opencv cascades to crop some positive samples from the video or better crop precisely the positive samples from your images. Negative samples are not a problem. Generate them randomly from your images. Clean and prepare data manually. This is the worst part. Crop also images selected correctly by default opencv cascade to achieve better results.
Second step
Build your initial detector, which is used for other positive data collecting. Again Clean and prepare data manually. Stupid boring and it takes a long time.
Third
!!Use this at your own risk. !!
Run the cloud machine on aws and install OpenCV 3.1.
Better is memory than a number of cores. Something like
r3.4xlarge cores 16 memory 52 $1.33 per Hour
!!!!On your own risk!!!!
8000 positive and 15000 negative datasets with 10- 20 stages could take minutes - hours instead of days of learning on your own computer. And it's cheaper than the electricity bill. Time to learn depends on a number of data, parameters, selected machines, and many other things.I can not guarantee that it goes so fast with your data and parameters. I do it this way save time. I am pretty sure that your first detector will be terrible. In the cloud, you can learn 10 detectors a day. Which is great
Performance detectMultiScale and parameters
- The source image is 640x480 image. My PC is a notebook I7, 4 cores, 8 threads, and a huge amount of RAM :). Opencv version is 3.1 default for Windows machines.
- I have 2 people detectors detectMultiScale in the main loop.
- The main performance issue is in the scaleFactor parameter and minSize of the object. In the faster case, I have 1.1, and in the slower case is 1.02 scaleFactor.
- Choose appropriately as needed the minSize parameter. If you have to small a window, your program needs to check far more options. Size(40,70)
CascadeClassifier at approximately 13 FPS
detectorBody.detectMultiScale(img, human, 1.1, 1, 0 | 1, Size(40,70), Size(80, 300));
detectorUpper.detectMultiScale(img, upperBody, 1.1, 1, 0 | 1, Size(40, 70), Size(80, 300));
CascadeClassifier at approximately 2 FPS
detectorBody.detectMultiScale(img, human, 1.02, 2, 0 | 1, Size(40,70), Size(80, 300));
detectorUpper.detectMultiScale(img, upperBody, 1.02, 2, 0 | 1, Size(40, 70), Size(80, 300));
Opencv HAAR LBP cascade download
I also learned some mine cascade on my own datasets. They are here. To be sure that you can reach them here are the links.
Cascade for car detector download
This is just a basic 5 stage haar cascade car detector, post where to find my cascade to detect cars
Cascade for head and people download
This cascade is also able to use by the tutorial code below. For head and for people detection.
Learned by me on my own data set. cascade for head and people detection
Opencv CascadeClassifier tutorial code
!!Simple as OpenCV Sample. Let me add some useful remarks. Some Windows OpenCV versions detect multiScale return errors.. 2.9.4 -3.0 maybe 3.1. I don't know why. The classifier returns to many output rectangles out of bound. This does not make sense and kills your program.
If this is an issue of your error on Windows. Switch from debug to release. It helps. !! Really
If this is an issue of your error on Windows. Switch from debug to release. It helps. !! Really
#include "opencv2\highgui.hpp"
#include "opencv2\imgproc.hpp"
#include "opencv2\objdetect\objdetect.hpp"
#include "opencv2/video/tracking.hpp"
#include <vector>
#include <stdio.h>
#include <Windows.h>
#include <iostream>
#include <time.h>
#include <ctime>
using namespace cv;
using namespace std;
int main(int argc, const char **argv)
{
// prepare video input
VideoCapture cap("input.mov");
// prepare video output
VideoWriter outputVideo;
outputVideo.open("video4.wmv", CV_FOURCC('W', 'M', 'V', '2'), cap.get(CV_CAP_PROP_FPS), Size(640, 480), true);
// prepare cascadeClassifier
CascadeClassifier detectorBody;
CascadeClassifier detectorUpper;
// !! Put your cascade or opencv cascede into project folder !!
string cascadeName1 = "cascadeName.xml";
string cascadeName2 = "cascadeName.xml";
// Load cascade into CascadeClassifier
bool loaded1 = detectorBody.load(cascadeName1);
bool loaded3 = detectorUpper.load(cascadeName2);
// Basic video input loop
for (;;)
{
bool Is = cap.grab();
if (Is == false)
{
cout << "Video Capture Fail" << endl;
break;
}
else
{
// Just for measure time
const clock_t begin_time = clock();
// Store results in these 2 vectors
vector<Rect> human;
vector<Rect> upperBody;
// prepare 2 Mat container
Mat img;
Mat original;
// capture frame from video file
cap.retrieve(img, CV_CAP_OPENNI_BGR_IMAGE);
// Resize image if you want with same size as your VideoWriter
resize(img, img, Size(640, 480));
// Store original colored image
img.copyTo(original);
// color to gray image
cvtColor(img, img, CV_BGR2GRAY);
// detect people, more remarks in performace section
detectorBody.detectMultiScale(img, human, 1.1, 2, 0 | 1, Size(40, 70), Size(80, 300));
detectorUpper.detectMultiScale(img, upperBody, 1.1, 2, 0 | 1, Size(40, 70), Size(80, 300));
// Draw results from detectorBody into original colored image
if (human.size() > 0)
{
for (int gg = 0; gg < human.size(); gg++)
{
rectangle(original, human[gg].tl(), human[gg].br(), Scalar(0, 0, 255), 2, 8, 0);
}
}
// Draw results from detectorUpper into original colored image
if (upperBody.size() > 0)
{
for (int gg = 0; gg < upperBody.size(); gg++)
{
rectangle(original, upperBody[gg].tl(), upperBody[gg].br(), Scalar(255, 0, 0), 2, 8, 0);
}
}
// measure time as current - begin_time
clock_t diff = clock() - begin_time;
// convert time into string
char buffer[126];
sprintf(buffer, "%d", diff);
// display TIME ms on original image
putText(original, buffer, Point(100, 20), 1, 2, Scalar(255, 255, 255), 2, 8, 0);
putText(original, "ms", Point(150, 20), 1, 2, Scalar(255, 255, 255), 2, 8, 0);
// draw results
namedWindow("prew", WINDOW_AUTOSIZE);
imshow("prew", original);
// make video output
outputVideo << original;
int key1 = waitKey(20);
}
}
}
I will try this code on my own. Hope it will help me.
Splendid Skyline Review
Awesome! Thanks for sharing code!!
where is the complete code? I haven't seen .hpp files
Hi. I am referening this tutorial. Could you give me this code? I want to reference it. Thanks a lot. My email: tiennt@mta.edu.vn
This comment has been removed by the author.
Hi! What an interesting blog!!
I'm working on a ROS module for person detection and I was wondering if you could give me the CascadeClassifier(body detector) you used because I couldn't find the link in this tutorial.
Thanks for your help!
My email: marisofia.iglesias@gmail.com
Hi,
Very good example and useful for some use cases. It is possible to send the code? My email: madeirajp@gmail.com
Thanks and good work
^
^
nice nice
sbobet
maxbet
แทงบอลออนไลน์
Good articles, well presented and informative articles.
แทงบอล sbobet
sbobet mobile
รับแทงบอล
Wow, use, I have to admit that what I see. There is a very interesting story.
แทงบอล maxbet
แทงบอล maxbet
ทางเข้า maxbet
This comment has been removed by the author.
Hi ad! In step 1, you mean using default .xml file of openCV to crop and get location for positive images, right ? And this solution is work for detecting human in different position like 'sitting'? thank you very much!!
This comment has been removed by the author.
Thanks for your help.
could you give me this code? i'm student, i want to reference it.
My email: "gpham378@gmail.com"
Thanks for your help.
could you give me this code? i'm student, i want to reference it.
My email: "gpham378@gmail.com"
This comment has been removed by the author.
Thanks a lot for such a nice tutorial....can i get this code of people tracking...i'm a student..my id aratimcs@gmail.com....Thanks
Excellent blog I visit this blog it's really awesome. The important thing is that in this blog content written clearly and understandable. The content of information is very informative.
Oracle Fusion HCM Online Training
Oracle Fusion SCM Online Training
Oracle Fusion Financials Online Training
Big Data and Hadoop Training In Hyderabad
oracle fusion financials classroom training
Workday HCM Online Training
Oracle Fusion HCM Classroom Training
Workday HCM Online Training
can you send me complete code?
rimsha479@gmail.com
Excellent blog
UFABET