Research on the Application of Visual SLAM in Embedded GPU

Algorithm 2

Optimization of feature detection algorithms and nonmaximum suppression algorithm processes.

Input: Image Gaussian pyramid Collection d_I_p[0…levelNum], Gaussian pyramid layers levelNum

Output: A set of feature points

1 Perform feature detection and nonmaximum suppression alternately

2 for i 1 to levelNum

3  gpuFeatureDetect(d_I_p[i]);

4  P_t[i-1][0…N_t2] nonmaximumSuppression(P_t[i-1][0…N_t1]);

5  P_t[i][0…N_t1] gpuGetFeaturePoint();

6 end

7 P_t[levelNum][0…N_t2] nonmaximumSuppression(P_t[levelNum][0…N_t1]);

Note: The beginning of d refers to the data allocated in the video memory, N_t1 and N_t2 are the temporary number of feature points