Ratio Based Distance Measurement for Automatic Pick and Place Robot.
K.M.I.Yasar Arafath, Adeep Shafi, Adnan Aboobacker,
Asha K V, Muhammed Mousoof K, Rashida P
P
ashakandanakam@gmail.com,
rashrashida11@gmail.com
Abstract— Automatic pick and place robots have wide applications
especially in the field of agriculture, industry, etc. The most important part
in the pick and place robots is the calculation of distance between the object
and the observer. Since cameras are becoming an integral part of the robot,
this work uses a camera for the measurement of the distance. The previous works
in this field either uses external hardware in addition to camera in order to
measure the distance or has knowledge regarding the size of the object. This
also makes the work useful in cases where measuring instruments cannot be used.
This work uses a ratio based distance measurement method for calculating the
distance between the observer and the object. The camera on the robot takes the
picture of the object and calculates the pixel height. Then the robot moves a
known distance towards the object and again calculates the pixel height of the
object by capturing a new image. The actual distance to the object is
calculated using the ratio of the pixel heights and the known distance moved by
the robot. Then robot moves the measured distance, picks the object and places
the object at the starting position. The MATLAB is used as the processing
software and Arduino Uno is interfaced with MATLAB to control the robotic
actions.
II.
LITERATURE SURVEY
There are
several methods for finding the distance to an object. In this section we
discuss three measuring principles. One is the traditional triangular measuring
method; second one is based on Fast Fourier Transform and comparison with data
stored in the database as reference; third method uses an equation for distance
calculation.
Dk = The distance between the two projected points
From
(1) Hk can be
calculated. The difference between Hmax and Hk gives the distance from camera to object.
D0= Distance to object, f= Focal length (mm)
Hro= Real
height of object (mm),Hip= image
height (pixel)
Hop = Object height (pixel)
J. Phelawan et al. in their paper ‘A new technique for distance measurement of between
vehicles to vehicles by plate car using image processing’ [2] describes
another method based on FFT. In this method camera can capture images and the
picture process is done by Fast Fourier Transform (FFT). The systems can be
compared with data stored in the database as reference. This method will be applicable to only the
objects whose size is known.
III.
RATIO BASED DISTANCE
MEASUREMENT METHOD
In (2) all parameters are constant except
pixel height of the object. So we can reduce as
d α 1/h (3)
d/(d-1)=h'/h (4)
where,
The
block diagram for automatic pick and place robot using camera is as shown in
fig. 2. This contains mainly eight blocks.
Camera is used as sensor
to measure the distance to the object. In modern robots camera is used as a
sensor since it scans the entire environment and can be used instead of many
other sensors for distance measurement, inspection, identification etc.
Lighting is
an essential part of machine vision. Good illumination of the scene is
important because of its effect on the level of complexity of image processing
algorithm. Poor lighting makes the task of interpreting the scene more
difficult. Proper lighting technique should provide high contrast and minimize
reflections and shadows.
Now the entire noise
will be removed as shown in Fig. 4. The number of rows containing white pixels
gives the pixel height of the object.
The ratio of pixel
heights of the object obtained from two positions will give us the distance to
the object using Eq. 4.
The Arduino Uno board is interfaced with
MATLAB. A server program is loaded to the arduino that will execute the
operations or commands requested by MATLAB and returns the datas needed. So, by
using MATLAB we can simultaneously control the arduino and camera.
[1] Ming-Chih Lu, Wei-Yen Wang and Chun-Yen Chu."Image-Based Distance and Area Measuring Systems”IEEE sensors
journal, vol. 6, no. 2, April 2006
Ratio Based Distance Measurement for Automatic Pick and Place Robot.
K.M.I.Yasar Arafath, Adeep Shafi, Adnan Aboobacker,
Asha K V, Muhammed Mousoof K, Rashida P
P
ashakandanakam@gmail.com,
rashrashida11@gmail.com
Abstract— Automatic pick and place robots have wide applications
especially in the field of agriculture, industry, etc. The most important part
in the pick and place robots is the calculation of distance between the object
and the observer. Since cameras are becoming an integral part of the robot,
this work uses a camera for the measurement of the distance. The previous works
in this field either uses external hardware in addition to camera in order to
measure the distance or has knowledge regarding the size of the object. This
also makes the work useful in cases where measuring instruments cannot be used.
This work uses a ratio based distance measurement method for calculating the
distance between the observer and the object. The camera on the robot takes the
picture of the object and calculates the pixel height. Then the robot moves a
known distance towards the object and again calculates the pixel height of the
object by capturing a new image. The actual distance to the object is
calculated using the ratio of the pixel heights and the known distance moved by
the robot. Then robot moves the measured distance, picks the object and places
the object at the starting position. The MATLAB is used as the processing
software and Arduino Uno is interfaced with MATLAB to control the robotic
actions.
Keywords— robotics, camera, pixel height, ratio based distance
measurement, MATLAB, Arduino Uno
I.
Introduction
Due to the shortage of man power, machines
are more used in the field of
agriculture. Automatic pick and place robots have wide applications, especially
in the field of agriculture, industry, etc. In the field of agriculture, the
fruit picking is an important process. Robots are used to identify the fruits
and calculate the distance between the robot and fruit. The robot moves towards
the fruit and collect it. In industries, the materials have to be collected and
moved from one area to another. Pick and place robots are becoming more common
in this field. So there is a great
significance in the distance calculation between the robot and the object. Once
we calculated the distance, we can do any operation using the same robot by
changing their end effectors. According to the application, the structure of
end effectors may differ, sometimes we have to grasp or pluck something.
Traditionally, there are several solutions for measuring distance. For
distance measurement, they can roughly be separated into two types, contact and
noncontact. In the case of contact measurement, there are lots of products to
measure distance. The chief defect is that the objects can be corrosive. For
noncontact measurement, there are different solutions, like laser reflection
and ultrasonic (or light) reflection. In the case of the laser techniques,
laser projectors and receivers are needed. Although using laser technology has
the advantage of speed, object reflectivity plays an important role. If the
object reflectivity is bad, the system will work poorly or not at all. The
ultrasonic technique also has the same problem caused by object reflectivity.
Since vision is becoming an integral part
of robots, in this work, the measurement of distance between the observer and
the object is done using camera. The previous works in this field make use of
external hardware and some knowledge about object size for distance measurement.
In some methods we also use focal length of the camera. This work does not use
any hardware other than camera for the distance measurement. The Ratio based
method is not depending upon focal length, external hardware and size of the
object, and only depend upon the known distance that the camera moved and pixel
heights of object before and after movement. MATLAB is one of the software
available which provide the platform for finding pixel heights.
Section II describes about the literature
survey. Followed that, ratio based distance measurement is explained and the
system description is given in the section IV.
II.
LITERATURE SURVEY
There are
several methods for finding the distance to an object. In this section we
discuss three measuring principles. One is the traditional triangular measuring
method; second one is based on Fast Fourier Transform and comparison with data
stored in the database as reference; third method uses an equation for distance
calculation.
Ming- Chih Lu et al. in their paper “ Image-Based Distance and Area Measuring Systems” [1] proposes a triangular method for measuring
the distance between the observer and the object. Here they use a CCD camera
and two laser projectors. Two
lasers are placed on the bottom and top of the CCD camera and the laser beams
produce a triangle as shown in the Fig. 1. Assume two
projected points on the two sides of the triangle.
Dk = The distance between the two projected points
Dmax = Distance between the two laser projects
Hk = Distance from two projected points to apex of tringle
Hmax = Distance from the CCD camera to apex of
the triangle
Dmax and Hmax are known to us. Dk can be
calculated by using an extra oscillator. An external clock generated by an
extra oscillator measure the time interval between two projected points.
External clock value is counted between two projected points. Actual distance Dk is proportional to external clock value.
From
(1) Hk can be
calculated. The difference between Hmax and Hk gives the distance from camera to object.
In
another method, the distance between a camera and an object is calculated by
using equation,
D0= Distance to object, f= Focal length (mm)
Hro= Real
height of object (mm),Hip= image
height (pixel)
Hop = Object height (pixel)
Hs= sensor height (mm)
Using this
equation we can measure the distance to an object. Any programming software
like Matlab can be used for this.
The disadvantage is that, it is necessary to know the real height of the
object. Otherwise we can’t proceed through this method. Measurement of the
object’s height is not a practical.
J. Phelawan et al. in their paper ‘A new technique for distance measurement of between
vehicles to vehicles by plate car using image processing’ [2] describes
another method based on FFT. In this method camera can capture images and the
picture process is done by Fast Fourier Transform (FFT). The systems can be
compared with data stored in the database as reference. This method will be applicable to only the
objects whose size is known.
J. Phelawan et al. in their paper ‘A new technique for distance measurement of between
vehicles to vehicles by plate car using image processing’ [2] describes
another method based on FFT. In this method camera can capture images and the
picture process is done by Fast Fourier Transform (FFT). The systems can be
compared with data stored in the database as reference. This method will be applicable to only the
objects whose size is known.
III.
RATIO BASED DISTANCE
MEASUREMENT METHOD
In (2) all parameters are constant except
pixel height of the object. So we can reduce as
d α 1/h (3)
where,
d= distance away from the object
h= pixel
height of the object
The camera captures the object from two
positions. The ratio of pixel heights of the object obtained from two positions
will give us the distance to the object, if the distance between two positions
is known. Initially the camera is at d
distance away from the object. The pixel height of object at that time is given
by h. Then robot moves a known
distance. Let it be 1cm. At that time the angle of camera is not changing, but
pixel height of object is varied to h'. So,
d/(d-1)=h'/h (4)
where,
d = actual distance to
object from camera
h = pixel height of object when camera at d
distance
h' =pixel height of object when camera at d-1
distance
The
block diagram for automatic pick and place robot using camera is as shown in
fig. 2. This contains mainly eight blocks.
Camera is used as sensor
to measure the distance to the object. In modern robots camera is used as a
sensor since it scans the entire environment and can be used instead of many
other sensors for distance measurement, inspection, identification etc.
Lighting is
an essential part of machine vision. Good illumination of the scene is
important because of its effect on the level of complexity of image processing
algorithm. Poor lighting makes the task of interpreting the scene more
difficult. Proper lighting technique should provide high contrast and minimize
reflections and shadows.
MATLAB is interfaced with camera using the
winvideo adapter. It helps us to take video input to the matlab and snapshot
required for processing is obtained.
To get the distance, we have to find out
the pixel height of object from the image. For that the object has to be
recognized from the image. The object is recognized by comparing the RGB values
to a threshold value. And the image will be converted to binary mode. Then the
object will be in white colour and the surrounding will be in black colour(Fig.
3).
The
next step is to remove the noises in the recognized image. The morphological
operations are used for noise rejection. The structural element used along with
morphological operations was a disk.
Now the entire noise
will be removed as shown in Fig. 4. The number of rows containing white pixels
gives the pixel height of the object.
The ratio of pixel
heights of the object obtained from two positions will give us the distance to
the object using Eq. 4.
The Arduino Uno board is interfaced with
MATLAB. A server program is loaded to the arduino that will execute the
operations or commands requested by MATLAB and returns the datas needed. So, by
using MATLAB we can simultaneously control the arduino and camera.
L293D is a dual H-bridge
motor driver integrated circuit (IC)[7]. Geared DC motors can be defined as an
extension of DC motor[8]. A geared DC Motor has a gear assembly attached to the
motor. Here, we use a12V, 60 rpm motor. So in one minute there is 60 rotations
and in one second there is one revolution.
Distance moved in
1s= circumference of wheel
= 2*pi* r =31.4 cm
Where,
r (radius of wheel) =5 cm
So, the obtained distance will be divided
by 31.4 and the robot will be allowed to move that much time to reach the
object. Now the arm motor will be turned on unless the limit switch connected
on the gripper is closed. And the robot moves back to the original position by
rotating the motor in reverse direction.
The sequences of operations are shown in
the flow chart below in fig. 5.
[1] Ming-Chih Lu, Wei-Yen Wang and Chun-Yen Chu."Image-Based Distance and Area Measuring Systems”IEEE sensors
journal, vol. 6, no. 2, April 2006
[2] J. Phelawan, P. Kittisut, N. Pornsuwancharoen.” A new technique for distance measurement of between vehicles to
vehicles by plate car using image processing”-science direct2011
[3] Mikell P Groover, Mitchell Weiss, Roger N Nagel, Nicholas G Odrey,
"Industrial Robotics: Technology,
Programming and Applications", McGRAW-HILL international edition
[4] Rafael C. Gonzalez and Richard E. woods, ‘Digital Image Processing’
[5] AnkurAgraval, “An elementary introduction toimage processing
based robot", IITkanpur
[6] http://www.arduino.cc/
[7] http://www.engineersgarage.com/electronic-components/l293d-motor-driver-ic
[8] http://www.engineersgarage.com/insight/how-geared-dc-motor-works
No comments:
Post a Comment