更新 Trace/handeye_calibration
This commit is contained in:
@ -1,4 +1,50 @@
|
||||
import numpy as np
|
||||
from scipy.spatial.transform import Rotation as R
|
||||
|
||||
|
||||
def flip_coefficient_if_positive(coefficient):
|
||||
# 检查 coefficient[2] 是否大于0
|
||||
if coefficient[2] > 0:
|
||||
# 取反所有分量
|
||||
coefficient = [-x for x in coefficient]
|
||||
print("翻转:\n")
|
||||
|
||||
return coefficient
|
||||
|
||||
def vec2ola(coefficient):
|
||||
coefficient_raw = flip_coefficient_if_positive(coefficient)
|
||||
coefficient = coefficient_raw.reshape(-1)
|
||||
curZ = np.array(coefficient)# 定义 Z 方向的向量
|
||||
# print(curZ)
|
||||
|
||||
curX = np.array([1, 1, 0],dtype=np.float64)# 定义初始 X 方向的向量
|
||||
curX /= np.linalg.norm(curX) # 归一化 X 向量
|
||||
|
||||
curY = np.cross(curZ, curX)# 计算 Y 方向的向量
|
||||
curY /= np.linalg.norm(curY) # 归一化 Y 向量
|
||||
|
||||
curX = np.cross(curY, curZ)# 重新计算 X 方向的向量
|
||||
curX /= np.linalg.norm(curX) # 归一化 X 向量
|
||||
|
||||
# 创建旋转矩阵
|
||||
rotM = np.array([
|
||||
[curX[0], curY[0], curZ[0]],
|
||||
[curX[1], curY[1], curZ[1]],
|
||||
[curX[2], curY[2], curZ[2]]
|
||||
])
|
||||
|
||||
# 打印旋转矩阵
|
||||
print("Rotation Matrix:")
|
||||
print(rotM)
|
||||
|
||||
# 计算欧拉角(XYZ顺序)并转换为度
|
||||
r = R.from_matrix(rotM)
|
||||
euler_angles = r.as_euler('xyz', degrees=True)
|
||||
|
||||
print("Euler Angles (degrees):")
|
||||
print(euler_angles)
|
||||
|
||||
|
||||
|
||||
def vec2attitude(a,b,c):
|
||||
pi = np.arccos(-1.0) # pi 的值
|
||||
@ -23,13 +69,14 @@ def vec2attitude(a,b,c):
|
||||
matrix_result = np.dot(np.dot(np.dot(r1, r2), r3), vector1)
|
||||
|
||||
# 输出结果
|
||||
rpy=[]
|
||||
ola=[]
|
||||
for m in range(matrix_result.shape[0]):
|
||||
rpy.append(matrix_result[m][0])
|
||||
ola.append(matrix_result[m][0])
|
||||
print(f"{matrix_result[m][0]:<20}")
|
||||
|
||||
return rpy
|
||||
return ola
|
||||
|
||||
#这个版本是先拿法向量转换成基坐标,在转欧拉
|
||||
#黄老师会给我传目标物的中心点坐标x,y,z和目标位姿的平面法向量a,b,c
|
||||
def getPosition(x,y,z,a,b,c):
|
||||
target = np.asarray([x, y, z])
|
||||
@ -37,16 +84,15 @@ def getPosition(x,y,z,a,b,c):
|
||||
position = np.dot(camera2robot[0:3, 0:3], target) + camera2robot[0:3, 3:]
|
||||
target_position = position[0:3, 0]#转换后的位置信息
|
||||
|
||||
# vector = np.asarray([a, b, c])
|
||||
# normal_vector = vector / np.linalg.norm(vector)#归一化
|
||||
# normal_vector.shape = (3, 1)
|
||||
# dot_angle = np.dot(camera2robot[0:3, 0:3], normal_vector)#转换后的法向量,方向依然是同一个方向,只是表示方法变了
|
||||
vector = np.asarray([a, b, c])
|
||||
normal_vector = vector / np.linalg.norm(vector)#归一化
|
||||
normal_vector.shape = (3, 1)
|
||||
dot_angle = np.dot(camera2robot[0:3, 0:3], normal_vector)#转换后的法向量,方向依然是同一个方向,只是表示方法变了
|
||||
|
||||
# angle_tool = R.from_matrix(camera2robot[0:3, 0:3])
|
||||
rpy = vec2attitude(a,b,c)
|
||||
r,p,y=rpy[0],rpy[1],rpy[2]#r表示u,p表示v,y表示w
|
||||
target_angle_raw = np.asarray([r,p,y])
|
||||
target_angle = np.dot(camera2robot[0:3, 0:3],target_angle_raw)
|
||||
print(target_position,target_angle)
|
||||
target_angle = vec2ola(dot_angle)#把转换之后的法向量转换为欧拉角,欧拉角不是rpy角
|
||||
|
||||
# r,p,y = angle_tool.as_euler('xyz',degrees=True)#r表示u,p表示v,y表示w
|
||||
# target_angle = np.asarray([r,p,y])
|
||||
# print(target_angle)
|
||||
|
||||
return target_position,target_angle
|
||||
Reference in New Issue
Block a user