算法2.点云簇相似性评分算法

输入：前后帧点云簇 $C_{\text{pre\_frame}}=\left(C_1,C_2,\cdots ,C_m\right)$ ， $C_{\text{cur\_frame}}=\left({C^{\prime }}_1,{C^{\prime }}_2,\cdots ,{C^{\prime }}_n\right)$ ；点云簇中心距离值 ${\text{dist}}_{\text{threshold}}$

输出：点云簇相似性评分 $\text{score}\left(C_i,{C^{\prime }}_j\right)$

1)for $o_i\in C_i$ in $C_{\text{pre\_frame}}$

2)for ${o^{\prime }}_j\in {C^{\prime }}_j$ in $C_{\text{cur\_frame}}$

3) if $\text{dist}\left(o_i,{o^{\prime }}_j\right)>{\text{dist}}_{\text{threshold}}$

4) ${\text{score}}_{\text{num}}\left(C_i,{C^{\prime }}_j\right)=0$

5) endif

6) ${\text{score}}_{\text{num}}\left(C_i,{C^{\prime }}_j\right)=\frac{\text{dist}\left(C_i,{C^{\prime }}_j\right)}{{\text{dist}}_{\text{threshold}}}$

7) ${\text{score}}_{\text{dist}}\left(C_i,{C^{\prime }}_j\right)=\frac{\min \left(\text{num}\left(C_i\right),\text{num}\left({C^{\prime }}_j\right)\right)}{\max \left(\text{num}\left(C_i\right),\text{num}\left({C^{\prime }}_j\right)\right)}$

8) $\text{score}\left(C_i,{C^{\prime }}_j\right)=\left(1-{\text{score}}_{\text{dist}}\left(C_i,{C^{\prime }}_j\right)\right)\ast \frac{1}{2}+{\text{score}}_{\text{num}}\left(C_i,{C^{\prime }}_j\right)\ast \frac{1}{2}$

9)endfor

10) endfor