Depth calculation of an Object Using Stereo Images

We calculate the depth of an object using stereo images with the help of two cameras and calibration because without calibration and two cameras we cannot get the accurate results.As per requirement we are doing this in nine degree freedom. So first we detect the end points of the object by

Project Title

Depth calculation of an Object Using Stereo Images

Project Area of Specialization

Cyber Security

Project Summary

We calculate the depth of an object using stereo images with the help of two cameras and calibration because without calibration and two cameras we cannot get the accurate results.As per requirement we are doing this in nine degree freedom.

So first we detect the end points of the object by using an algorithem that initially there is no vriation in front of the cameras when a variation come in front of the camera,variation would be detected by the camera and after variation where variation is ended so the cameras detect that point so these are our first two end points and then through linera interpolation we detect the remaining seven points and after that through tringulation we extracted the nine 3D points. After that we estimate the depth of the object by applying distnce farmula.

[We just use two "A4 TECH " cameras and cables.]

Project Objectives

Depth perception among living things is due to the presence of two eyes. Each eye forms a separate picture of the scene and transmits the image to brain to perceive depth. The same concept is applied in modern world devices to perceive depth.  To estimate depth from stereo images of various objects invariant to their orientations or the degrees of freedom. 

So our objective is to detect object from stereo images invariant to image quality and to estimate depth of the detected object invariant to its orientation.

Project Implementation Method

We just use matlab software.

As per requirement we are doing this in nine degree freedom.

So first we detect the end points of the object by using an algorithem that initially there is no vriation in front of the cameras when a variation come in front of the camera,variation would be detected by the camera and after variation where variation is ended so the cameras detect that point so these are our first two end points and then through linera interpolation we detect the remaining seven points and after that through tringulation we extracted the nine 3D points. After that we estimate the depth of the object by applying distnce farmula.

Benefits of the Project

Through this project we can get the accurate results in low cost.We can also calculated the depth using laser but it cost was high.

Second thing we can use this project in eroplane which tell us that how far away is that objectso we can esily choose our trget.

Technical Details of Final Deliverable

At first, the input frames are converted to grayscale.Afterwards, the gradients are computed in order to detect the object within the frame.Since the object is movable, therefore, it will have the maximum amount of change. The strength of the change is measured by computing the gradient magnitude.After computing the strength of change, it is used to detect the desired object.Different postprocessing steps (such as blob elimination and morphological skeletonization) are performed to enhance the detected object (rod) information.The location of the detected object is further utilized to limit the object search area within the consecutive frames.

After detecting the object (which is the rod in our case), the corner points of the object are extracted.In order to extract the corner points, the processed frames are transposed.From each column of the transposed frame, the first and last transition between foreground and background pixels is measured.From these transitions, the four corner points transitions are extracted.Using these corner points, the two extreme points of the rod are measured.After tht by piecewise interpolation we detect the remaining seven points and by tingulation we detect the 3D of all the nine points and estimate the depth.

Final Deliverable of the Project

HW/SW integrated system

Type of Industry

Security

Technologies

Others

Sustainable Development Goals

Decent Work and Economic Growth

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com