To obtain the transmission line information for the maintenance of high and ultrahigh voltage electrical network, some
methods of automatic power line and spacer extraction from high-resolution remote sensing images are presented in this
paper. Aiming at the features of power lines and spacers in the high-resolution remote sensing images, a local linear
detector is used to detect pixels of linear features in any directions. According to the responses of the detector, a score
image is constructed, and an optimum method is used to determine the threshold according to different background.
Then the pixels of power lines and spacers can be detected automatically from the score image. After thinning of the
pixels, a curve fitting method is applied to obtain segments of power lines and spacers. As the influence of many
disturbing factors, the detected segments are discontinuous, and there are even some false detections. Based on shape
information of the power lines and spacers, the correct segments are chosen, and they are connected into complete form
of power lines or spacers. The experimental results indicate that our methods can accurately extract power lines and
spacers with different shapes from the high-resolution remote sensing images.