cv2.perspectiveTransform()透视变换原理

opencv中透视变换

$$k\left[\begin{array}{c}{x}^{\prime }\\ y^{\prime }\\ 1\end{array}\right]=H\left[\begin{array}{c}x\\ y\\ 1\end{array}\right]$$

corners = np.array([[281, 238], [325, 297], [283, 330],[248, 325]], dtype=np.float32).reshape(-1, 1, 2)
H = np.array([1,0,0,0,1,0,1,0,1], np.float32).reshape(3, 3)
new = cv.perspectiveTransform(corners, H)
corners = np.array([281, 238,1], dtype=np.float32)
vec=np.matmul(H, corners)
print(vec[0]/vec[2],vec[1]/vec[2])
print(new[0][0])

注意

perspectiveTransform()输入至少需要4个点，且点的坐标矩阵为（-1，1，2）即n个一行两列的坐标矩阵，两列代表横纵坐标，其等价于将坐标齐次化后，与H矩阵的乘法运算

运行结果