A three-dimensional object detection device can enhance the accuracy in detecting a three-dimensional object regardless of the brightness in the detection environment. The device has an image capture means that captures an image of a predetermined area and an image conversion means that converts the image through a viewpoint conversion into birds-eye view image. A first three-dimensional object detection means aligns positions of bird's-eye view images at different times obtained by the image conversion means, counts the number of pixels that exhibit a predetermined difference on a differential image of the aligned bird's-eye view images to generate a frequency distribution thereby creating differential waveform information, and detects a three-dimensional object on the basis of the differential waveform information. A second three-dimensional object detection means detects edge information from the bird's-eye view image obtained by the image conversion means and detects a three-dimensional object on the basis of the edge information.