Crop row location in real time is often an important goal in
the autonomous guidance of agricultural vehicles.
However, in this context, the crop rows are roughly approximated by
lines, while for weed patch detection the precision required is much
higher. Some other works use the Hough transform , to fully locate the crop rows and then label the rest of vegetation pixels as weeds.
The drawback of this approach is the high computational complexity of
the Hough transform, which makes it unsuitable for applications in
which there is a need to process images in real-time, i.e. at 25 fps, the standard video camera frame rate.
Finally, some other studies deal with simpler images, taken closer to the
ground and in such a way that perspective is eliminated, so that
crop rows can be more easily located and the processing adapted
to real-time
Crop row location in real time is often an important goal inthe autonomous guidance of agricultural vehicles. However, in this context, the crop rows are roughly approximated bylines, while for weed patch detection the precision required is muchhigher. Some other works use the Hough transform , to fully locate the crop rows and then label the rest of vegetation pixels as weeds. The drawback of this approach is the high computational complexity ofthe Hough transform, which makes it unsuitable for applications inwhich there is a need to process images in real-time, i.e. at 25 fps, the standard video camera frame rate.Finally, some other studies deal with simpler images, taken closer to theground and in such a way that perspective is eliminated, so thatcrop rows can be more easily located and the processing adaptedto real-time
การแปล กรุณารอสักครู่..
