Update README.md

README.md

@@ -1,16 +1,50 @@
 # lidar-camera-fisheye-calibration
 
-this calibration code is based on pnp algorithm,you should manually selet feature pointcloud from pcd and correspondens feature pixel on image.you better select over 
-three pairs to get better result.
 
-if you want calibrate fisheye between lidar: just change distortion vector D in calib.py (k1,k2,k3,k4)
+this calibration code is based on pnp algorithm.
+
+our algorithm support calibration between (fisheye/pinhole)camera and lidar 
+
+We give two way to calibrate lidar and camera:
+
+1)calib.py
+
+you can manually selet pointcloud feature and image corner feature
+
+![image](https://user-images.githubusercontent.com/42079541/215989229-b936952a-116c-4696-9771-133aabd85228.png)
+![image](https://user-images.githubusercontent.com/42079541/215989345-e45a4015-1454-4cf0-8410-fe126b7b9093.png)
+
+you can change the number of feature point at code :for i in range(4):
+
+2)calib_multi.py
+
+you should store your image in folder: your_path/lidar_camera_calib/calib_data/img
+![image](https://user-images.githubusercontent.com/42079541/215990850-e83d1986-8e44-44ac-8912-83ab1fb511eb.png)
+
+
+you should store your pcd in folder: your_path/lidar_camera_calib/calib_data/pcd
+![image](https://user-images.githubusercontent.com/42079541/215990902-d05b2782-b907-4fb3-ad82-fcf2e87d6ae1.png)
+
+set cropped_flag = True,and then you can crop chessboard in pointcloud
+
+when pointcloud winodw is open,press k to make window stable and then press ctrl+left mouse to crop chessboard.
+
+![image](https://user-images.githubusercontent.com/42079541/215991733-24545882-e0b1-450d-afca-48dea7570139.png)
+after crop chessboard ,press s in window and save pcd like crop1.pcd,crop2.pcd,crop3.pcd...(attention:you should start from 1)
+
+after done this,you will get rotation vetor and translation vetor and rotation matrix between cam and lidar 
+![image](https://user-images.githubusercontent.com/42079541/215993428-53335830-cad8-4d36-aa55-08a4321e8461.png)
+
+# how to change camera intrinsic
+if you want calibrate fisheye(kb4 model) between lidar: just change distortion vector D dist_coeffs = np.array([[k1],[k2],[0],[0],[k3],[k4],[0],[0]])
 
 if you want calibrate pinhole betwwen lidar: just change distortion vector D in calib.py to D =[[k1],[k2],[k3],[p1],[p2]]
 
 you can calibrate rotation and translation between fisheye/pinhole and lidar by calib.py.
 
 I provide a test sample to let you try.
 
+# result
 below is reprojection result:
 
 ![fisheye1](https://user-images.githubusercontent.com/42079541/215985248-4bf2c664-d873-4305-9ffa-f648a70e0dae.jpg)