29-09-2016, 10:45 AM
1456732040-illakiya.pdf (Size: 1.21 MB / Downloads: 5)
Abstract — In this paper, A low-power Laser projector and a Webcam are used to design a monitor system for the body
breath detection, which monitors and records the patient's breath and sends the information to the server. Our system
consists of two parts in which the first part has a Webcam, which is used to capture the images of the reflection from a
low-power laser projector and transmits it to a PC. For the second part, an image processing program using the temporal
differencing algorithm is used to detect the reflective mark movement to determine the breath rate, installed in the PC.
The low-power laser which we are using in this project is safe for the patients.
INTRODUCTION
The traditional ways of monitoring the breath rate involve contact with the human body . For example, for Impedance
Pneumography, one side of the electrode, which is placed on the skin of the chest, sends a high-frequency current and
concurrently uses the other side of the receiving electrode to receive the current changes that take place during a breathing cycle.
The earlier methods require some sort of tying up or other contact with the body. Thus there is a need to design a method of
breath detection that avoids contact with the human body.
Our design detects chest expansion and contraction in a way that is similar to the detection of a moving object by laser
reflection . There are three major methods used to detect a moving object. First, “temporal differencing” is used to find the
difference between two continuous image data and obtains the changes in the volume of a moving object .If the background
variation is not great, this method works well. Second, “background subtraction” establishes the background of the images and
then inputs one image after another in order to cancel the background and to obtain the moving object .This method, which
produces a complete outline of moving objects, is very sensitive to changes in the environment because the background is
established from the very beginning. Third, “optical flow” calculates moving locations in time to discern each pixel in the image.
After comparing these three methods we have used the image processing method with “temporal differencing” in our
design . This method has a higher image processing speed and its use in the circuit board will not cause much of the complexity in
the circuit. To enhance the accuracy of our detection system we measure the breath twice instead of a single laser reflection. This
design provides more amplification of the chest contraction from the breath and reduces the system detection error.
II. OPERATION OF THE SYSTEM
Fig 1 indicates block diagram. First we use a low power laser irradiating a patient’s chest with a smooth reflection
surface and let the laser reflect this onto a mirror. Then two mirrors will reflect the laser ray twice onto the projection wall. Now
we use the Webcam to monitor the laser projection. For home healthcare our design sends images captured by the Webcam to the
embedded board which through digital image processing determines whether the breath rate is normal. The LCD displays the realtime
status of the breath rate on the embedded board. If no signs of breath are detected, a warning message is sent to the hospital
by the network. We use an embedded board instead of a PC because of its low power consumption and portability. In addition the
embedded board reduces the cost, and one can use this system easily at home. It is thus convenient for home health application.
USING THE METHODS
A. DIFFERENCING METHOD
The temporal differencing method which reads the continuous image from the Webcam, then fetches the separated image
and processes it at the gray level. This method finds the subtraction of the continuous images, chooses the threshold value for the
decision, uses a Gaussian filter to remove noise and then obtains the moving object.
B. MAGNIFICATION METHOD
When a patient breathes, the chest displacement is small. So we propose the laser reflection method to enlarge the
displacement of the chest to detect the breath rate more easily[4]. Fig.2 shows the projection length of the moving object. X is the
downward displacement of the chest which can be amplified by laser reflection onto a tilted wall. Y is the projection length of the
first reflection. The magnification of the first reflection is shown in Eqn. (1).
?
?
= 2×(sin? ?(? + ?)+ ?) (1)
Eqn.(1) shows the relationship between magnification and oblique angle of the wall which can provide a 10-times amplification if
? ≤ 75° and ? + ? ≥ 80°. When ? + ? ≥ 80°, the projecting point on the wall will be so far away that the laser ray will generate
some distortion. The total magnification of the laser reflection is shown in Eqn. (2).
?
?
= 2×(sin? ?(? + ?)+ ?)(sin? ?(? + ? + ?) + ?) (2)
By using Eqn.(1) and Eqn. (2) we learn that the total magnification is the multiplication of the first and second reflection’s
amplification. To avoid distortion of the laser projection we arrange the first reflection to be enlarged by about 4.5 times
when (? + ?) ≥ 70° . Then we arrange ? ≥ 10° so that the second reflection can be enlarged more than 2.2 times. Therefore we
reach our goal of the laser projection being enlarged by more than 10 times with less distortion. Fig 2 indicates Projection length
of the chest displacement.
BREATH MONITORING
The subtraction of two continuous images results in different values when read in a different order of images. It is stored
in the matrix, and the elements are reset to zero if this subtraction is negative, because the size of the 8-bit gray scale element is 0-
255. The human body’s complete breath consists of the chest rising and falling, from maximum to minimum. If the chest rises to
the maximum and then not only falls but also instantaneously changes direction, this development may cause a turning point.
Three turning points are a complete breath cycle, and from them we calculate the breath rate. In addition we use our digital
method to enlarge the body movement to increase the detection distance. Fig 3 is the Flowchart of breath rate detection and it also determines the breath abnormalities. When the system detects no breath for more than 10 seconds, the system sends out a
warning message. The judging method decides whether the patient’s condition is normal or not by detecting the breath rate.
CONCLUSION
Our design uses the low-power laser reflection method to spot the variation in the breath and Webcam is used to capture
the image, the temporal differencing method for breath detection and an embedded board to monitor the patient’s breath without
the inconvenience to the body because of any physical contact by the instrument. This design detects chest expansion and
contraction more easily. The additional cost is small for monitoring a patient’s breath in home healthcare.