创新实训第六周总结

上周我得到中值坐标后，下一步是求解f(x)

相应代码

k=32 #防跳跃系数 
for i in range(1,len1):
    for j in range(len2):
        
#         before,min_value = find_min(j,dp_matrix_value[i-1,j-k:j+1].tolist()[0],k)
        before,min_value = find_min(j,get_list(dp_matrix_value[i-1,:],j-k,j+1),k,len2)
        dp_matrix_value[i,j] = min_value + distance_2d(data1[i,0],data1[i,1],data2[j,0],data2[j,1])
        dp_matrix_father[i,j] = before
    print(i,"/",len1)
        
        
# for i in range(len1):
    
#     temp = dp_matrix_value[len1-1-i,:].tolist()[0]
#     minV = min(temp)
#     pos = temp.index(minV)
#     print(len1-1-i,"->",dp_matrix_father[len1-1-i,pos])
 
temp = dp_matrix_value[len1-1,:].tolist()[0]
minV = min(temp)
pos = temp.index(minV)
print(len1-1,"->",pos)
 
 
# pos = dp_matrix_father[len1-2,pos]
# print(pos)
# pos = dp_matrix_father[len1-3,int(pos)]
# print(pos)
 
for i in range(len1-1):
    j = dp_matrix_father[len1-2-i,int(pos)]
    print(len1-2-i,"->",j)
    pos = j

可视化轮廓匹配度

其中，黄色的为原图，蓝色的是映射后的，可以看到而这匹配度还是很高的

可见对应关系还是比较准确的。