50
2 RFID System Physical Anti-Collision Experimental Verification
Fig. 2.12 The scale space of
SURF algorithm
matrix is obtained by using the integral image of the original graph, and the extremum
points of different scales are obtained. The scale space of SURF algorithm can be
shown in Fig. 2.12.
Similar to SIFT algorithm, each pixel point is compared with 26 points in its
neighborhood, leaving a maximum point. If the maximum point is less than the
threshold of the Hessian determinant, this point will be excluded.
(2) Describe feature point
After finding the feature points, the next step is to determine the direction of the
feature points. In SURF algorithm, feature points are taken as the center of a circle
and a circular neighborhood is determined. A 60-degree sector area is selected every
0.2 radians, and all points of small porter’s feature are counted. The direction with
the most points is the main direction of the feature point, as shown in Fig. 2.13. After
determining the main direction of feature points, the image is divided into 4 × 4 blocks
along the main direction, and then the Haar wavelet template is used to calculate each
block, and the statistical calculation result is the feature vector of feature points.
Fig. 2.13 Determine the main direction
2 RFID System Physical Anti-Collision Experimental Verification
Fig. 2.12 The scale space of
SURF algorithm
matrix is obtained by using the integral image of the original graph, and the extremum
points of different scales are obtained. The scale space of SURF algorithm can be
shown in Fig. 2.12.
Similar to SIFT algorithm, each pixel point is compared with 26 points in its
neighborhood, leaving a maximum point. If the maximum point is less than the
threshold of the Hessian determinant, this point will be excluded.
(2) Describe feature point
After finding the feature points, the next step is to determine the direction of the
feature points. In SURF algorithm, feature points are taken as the center of a circle
and a circular neighborhood is determined. A 60-degree sector area is selected every
0.2 radians, and all points of small porter’s feature are counted. The direction with
the most points is the main direction of the feature point, as shown in Fig. 2.13. After
determining the main direction of feature points, the image is divided into 4 × 4 blocks
along the main direction, and then the Haar wavelet template is used to calculate each
block, and the statistical calculation result is the feature vector of feature points.
Fig. 2.13 Determine the main direction