System and method for obstacle avoidance

Abstract

A method for acquiring an obstacle distance including determining a detection mode for detecting an obstacle distance of an obstacle; detecting the obstacle distance using the detection mode. Detecting the obstacle distance includes in response to determining a monocular mode as the detection mode: capturing a first image and a second image using a lens of an imaging device at two different times with a predetermined monocular imaging interval at two different locations; and calculating the obstacle distance via a monocular triangulation based on the first image, the second image, and a displacement of the imaging device between the two different times. Both the first image and the second image contain the obstacle. The predetermined monocular imaging interval varies based upon an altitude of the imaging device changes.

What is claimed is: 
     
         1 . A method for acquiring an obstacle distance, comprising:
 determining a detection mode for detecting an obstacle distance of an obstacle;   detecting the obstacle distance using the detection mode, including in response to determining a monocular mode as the detection mode:
 capturing a first image and a second image using a lens of an imaging device at two different times with a predetermined monocular imaging interval at two different locations, both the first image and the second image containing the obstacle, where the predetermined monocular imaging interval varies based upon an altitude of the imaging device changes; and 
 calculating the obstacle distance via a monocular triangulation based on the first image, the second image, and a displacement of the imaging device between the two different times. 
   
     
     
         2 . The method of  claim 1 , further comprising:
 determining the displacement of the imaging device between the two different times via an inertial measurement unit (IMU).   
     
     
         3 . The method of  claim 1 , wherein determining the detection mode includes:
 selecting the detection mode from a plurality of detection modes including the monocular mode and a binocular mode.   
     
     
         4 . The method of  claim 3 , wherein selecting the detection mode includes:
 obtaining a disparity between two images of the obstacle; and   selecting the detection mode from the plurality of detection modes based on the disparity between the two images.   
     
     
         5 . The method of  claim 4 , wherein selecting the detection mode further includes:
 comparing the disparity with a predetermined disparity threshold level; and   selecting the monocular mode as the detection mode in response to the disparity being less than the predetermined disparity threshold level, and selecting the binocular mode as the detection mode in response to the disparity being greater than or equal to the predetermined disparity threshold level.   
     
     
         6 . The method of  claim 4 ,
 wherein obtaining the disparity includes associating each feature point of a first one of the two images with a corresponding point of a second one of the two images; and   the method further comprising:
 generating a three-dimensional (3D) stereoscopic cloud of the obstacle based on the feature points and the obstacle distance. 
   
     
     
         7 . The method of  claim 3 , wherein detecting the obstacle distance further includes in response to selecting binocular mode as the detection mode:
 simultaneously capturing a third image with a first lens of the imaging device and capturing a fourth image with a second lens of the imaging device;   determining a baseline between the first lens and the second lens; and   calculating the obstacle distance using a binocular triangulation based on the third image, the fourth image, and the baseline.   
     
     
         8 . An unmanned aerial vehicle (UAV), comprising:
 a plurality of propellers for providing a lifting power; and   one or more processors, individually or collectively, configured to:
 determine a detection mode for detecting an obstacle distance of an obstacle; 
 detect the obstacle distance using the detection mode, including in response to determining a monocular mode as the detection mode:
 capturing a first image and a second image using a lens of an imaging device at two different times with a predetermined monocular imaging interval at two different locations, both the first image and the second image containing the obstacle, where the predetermined monocular imaging interval varies based upon an altitude of the imaging device changes; and 
 calculating the obstacle distance via a monocular triangulation based on the first image, the second image, and a displacement of the imaging device between the two different times. 
 
   
     
     
         9 . The UAV of  claim 8 , further comprising an inertial measurement unit (IMU) configured to:
 determine the displacement of the imaging device between the two different times.   
     
     
         10 . The UAV of  claim 8 , wherein the one or more processors are further configured to:
 select the detection mode from a plurality of detection modes including the monocular mode and a binocular mode.   
     
     
         11 . The UAV of  claim 10 , wherein the one or more processors are further configured to:
 obtain a disparity between two images of the obstacle; and   select the detection mode from the plurality of detection modes based on the disparity between the two images   
     
     
         12 . The UAV of  claim 11 , wherein the one or more processors are further configured to:
 compare the disparity with a predetermined disparity threshold level; and   select the monocular mode as the detection mode in response to the disparity being less than the predetermined disparity threshold level, and select the binocular mode as the detection mode in response to the disparity being greater than or equal to the predetermined disparity threshold level.   
     
     
         13 . The UAV of  claim 11 , wherein the one or more processors are further configured to:
 associate each feature point of a first one of the two images with a corresponding point of a second one of the two images; and   generate a three-dimensional (3D) stereoscopic cloud of the obstacle based on the feature points and the obstacle distance.   
     
     
         14 . The UAV of  claim 8 , wherein the one or more processors are further configured to:
 control the UAV to avoid the obstacle based on the obstacle distance.   
     
     
         15 . The UAV of  claim 14 , wherein the one or more processors are further configured to:
 perform motion estimation of the imaging device; and   control the UAV to avoid the obstacle based on the obstacle distance and the motion estimation.   
     
     
         16 . The UAV of  claim 15 , wherein the one or more processors are further configured to:
 perform the motion estimation with a binocular mode or the monocular mode.   
     
     
         17 . The UAV of  claim 16 , wherein the one or more processors are further configured to:
 determine a motion transformation of the imaging device between different times; and   
     
     
         18 . The UAV of  claim 17 , wherein the one or more processors are further configured to:
 acquire a velocity, a rotation, and an orientation of the imaging device for determining the motion transformation.   
     
     
         19 . The UAV of  claim 18 , wherein further includes an IMU configured to acquire the velocity, the rotation, and the orientation of the imaging device. 
     
     
         20 . An apparatus for assisting obstacle avoidance of a mobile platform, comprising one or more processors, individually or collectively, operate to:
 determine a detection mode for detecting an obstacle distance of an obstacle;   detect the obstacle distance using the detection mode, including in response to determining a monocular mode as the detection mode:
 capturing a first image and a second image using a lens of an imaging device at two different times with a predetermined monocular imaging interval at two different locations, both the first image and the second image containing the obstacle, where the predetermined monocular imaging interval varies based upon an altitude of the imaging device changes; and 
 calculating the obstacle distance via a monocular triangulation based on the first image, the second image, and a displacement of the imaging device between the two different times.