Highway Monitoring System

         

                Highway Monitoring System

I have done a project which will monitor the highways traffic. it’s a pre-programmed system, if we install it on the highway roads, it will detects the number of cars that is passing and we can control the traffic lights and we can give some instructions to the travelers through display, it’s an embedded system that interface the computer vision to real time device like MCU, display and traffic light. We can further improve this project to detect the speed of the vehicle if vehicle exceeds the speed this system can note vehicle number and we can also monitor the toll gate etc.

Here in this project I just use the feature based cascade classifier to detect the vehicles. Haar cascade classifier is best and accurate feature based classifier to detect the particular objects, human beings, face etc.
 in this project i used the recorded video we can use the live stream video also.we need to import the classifier and need to write small logic to draw the shapes on the on the moving object that is recognized by HAAR classifier we can observe how through below image how the classifier is working here i took the rectangular shape around the detected object .

next we need to detect the vehicles and count them that are passing by the road. For that I wrote logic to find center point of the object and I draw a line on two roads so that if the center point of the vehicle passes by the line then count will increase.

Now it’s time to interface it with the MCU and operate the real time objects like traffic lights and display.

traffic lights:

adding display:

You’re always welcome to give me your valuable suggestion through comments, if you need more information regarding this please leave a comment so that I can answer you.

Face Detection

                            Face Detection

I have seen an article about the HAAR classifier .which inspires me to work on image processing. image processing nothing but convert the image and video to digital form, digital form to arrays by using these arrays with some logic we can perform some actions that can change the features of the image/video.

Haar classifier is the best and effective object detection method proposed by Paul Viola and Michael Jones. cascade function is  trained by positive and negative images.Haar classifier is developed using Haar wavelets  .this wave lets is similar to fourier transformations if you want to know more about the Haar wavelets and classifiers just open the links below:

Haar classifier: https://en.wikipedia.org/wiki/Haar-like_features

Haar wavelets: http://www.whydomath.org/node/wavlets/hwt.html

With the knowledge of these we can make or train our own Haar classifier.

Here in this project I try to detect the face, eyes for which I use the existing classifier I found it when I Google about the classifier. We can use four edge detection rather than the classifier but it is not as accurate as Haar classifier and object detection not easy in four edge detection so I opt for haar classifier.

Let’s start our project:

Here i used the recorded video,we can use the live stream video also.we need to import the classifier and need to write small logic to draw shape on the feature that is recognized by haar classifier we can observe through below image how this classifier is working .

You’re always welcome to give me your valuable suggestion through comments, if you need more information regarding this please leave a comment so that I can answer you.

Computer vision or image processing

         Computer vision or image processing

I have done a project which will recognize the moving objects using computer vision we can say image processing

In order to do this project we need to work on some examples like how to import the image/video and convert it to the digital form, digital form to arrays by using these arrays with some logic with required libraries we can perform some actions that can change the features of the image/video.

Let’s start:

First we need to know how to import the image/video and change the color and save image/video using open CV

Next we need to know how to draw shapes and write a text on the image /video

Now we need to learn how to decrease the noise in particular frame/image. Here I am using threshold, gaus , otus. See the difference from the original image to processed images

Now we need to work on finding the edges

We can crop the image in open CV

Now we need to detect the moving object

Finally we will detect the object by using above learned topics.here i try to detect the objects by four edge detections without using any cascading .