136
4 Image Theory of RFID System Physical Anti-Collision
∇
2 f (x, y) = f (x + 1, y) + f (x − 1, y) + f (x, y + 1) + f (x, y − 1) − 4f (x, y)
(4.37)
Since the Laplace operator is a second-order differential operator, it is very sensitive to noise in the image. The LOG operator belongs to a detection operator improved
based on the classic operator. The signal-to-noise ratio needs to be determined during
the calculation of the optimal filter.
The LOG operator has three main advantages. First, the calculation speed is
relatively fast; second, the Gaussian filter can achieve the best in the frequency
domain and the space domain; third, it has the excellent anti-jamming ability, good
continuity, strong accuracy, and can reduce the contrast. The boundary is accurately
presented. However, this method also has certain limitations. For example, if the
width of the operator is greater than the width of the boundary, the zero-crossing
slope will merge, which will cause the problem of losing some boundary details.
LOG = ∇
2 G(x, y) =
1
πσ 4
x
2
+ y
2
2σ 2 − 1
exp
−
x
2
+ y
2
2σ 2
(4.38)
In the process of detecting step edges by zero-crossing, the LOG operator is
the best operator at this stage, but in fact, the zero-crossing points are not all edge
points. So when determining the zero-crossing point, it is necessary to verify whether
it is accurate. In addition, in essence, the facet model is also a method for edge
detection based on second-order differential zero-crossings.
4.4 3D Space RFID Tag Location
The experimental environment of the three-dimensional RFID tag positioning in this
study is shown in Fig. 4.18. Figure 4.18a is the matching pattern attached to the bottom
of the tag holder. The whole positioning process is based on the three-dimensional
multi-tag positioning system described in Chap. 2.
First, the system will grayscale the acquired pictures, and the deblur process is
performed on the acquired image by the method of de-motion blur described in
Chap. 3. As shown in Fig. 4.18b, it is the image for matching positioning captured
by the vertical CCD camera after deblur process. The threshold segmentation of the
picture highlights the need to obtain the tag holder pattern and the center pattern of
the turntable, as shown in Fig. 4.19.
Then, the system determines the center position of the turntable based on the image
matching technology of the center of the turntable, as shown in Fig. 4.19b. At the
same time, the image matching of the tag holder pattern is also performed. The ratio
of each pixel in the image to the actual length can be determined from the vertical
distance between the turntable and the vertical CCD and the CCD camera itself and
the field of view size. The horizontal camera and vertical camera have been calibrated
4 Image Theory of RFID System Physical Anti-Collision
∇
2 f (x, y) = f (x + 1, y) + f (x − 1, y) + f (x, y + 1) + f (x, y − 1) − 4f (x, y)
(4.37)
Since the Laplace operator is a second-order differential operator, it is very sensitive to noise in the image. The LOG operator belongs to a detection operator improved
based on the classic operator. The signal-to-noise ratio needs to be determined during
the calculation of the optimal filter.
The LOG operator has three main advantages. First, the calculation speed is
relatively fast; second, the Gaussian filter can achieve the best in the frequency
domain and the space domain; third, it has the excellent anti-jamming ability, good
continuity, strong accuracy, and can reduce the contrast. The boundary is accurately
presented. However, this method also has certain limitations. For example, if the
width of the operator is greater than the width of the boundary, the zero-crossing
slope will merge, which will cause the problem of losing some boundary details.
LOG = ∇
2 G(x, y) =
1
πσ 4
x
2
+ y
2
2σ 2 − 1
exp
−
x
2
+ y
2
2σ 2
(4.38)
In the process of detecting step edges by zero-crossing, the LOG operator is
the best operator at this stage, but in fact, the zero-crossing points are not all edge
points. So when determining the zero-crossing point, it is necessary to verify whether
it is accurate. In addition, in essence, the facet model is also a method for edge
detection based on second-order differential zero-crossings.
4.4 3D Space RFID Tag Location
The experimental environment of the three-dimensional RFID tag positioning in this
study is shown in Fig. 4.18. Figure 4.18a is the matching pattern attached to the bottom
of the tag holder. The whole positioning process is based on the three-dimensional
multi-tag positioning system described in Chap. 2.
First, the system will grayscale the acquired pictures, and the deblur process is
performed on the acquired image by the method of de-motion blur described in
Chap. 3. As shown in Fig. 4.18b, it is the image for matching positioning captured
by the vertical CCD camera after deblur process. The threshold segmentation of the
picture highlights the need to obtain the tag holder pattern and the center pattern of
the turntable, as shown in Fig. 4.19.
Then, the system determines the center position of the turntable based on the image
matching technology of the center of the turntable, as shown in Fig. 4.19b. At the
same time, the image matching of the tag holder pattern is also performed. The ratio
of each pixel in the image to the actual length can be determined from the vertical
distance between the turntable and the vertical CCD and the CCD camera itself and
the field of view size. The horizontal camera and vertical camera have been calibrated