| 摘要 |
An objective of the present invention is to generate a more natural composite image wherein a solid object (obstacle) is easily seen. An image detection unit (106) extracts image feature values from images of each camera (101). An image conversion unit (105) computes a blend rate according to the image feature values and composites an image of a superposition area wherein a plurality of camera images overlap. An assessment is made of a correspondence of the image feature values of each image in the superposition area, and a determination is made that a solid object is present if the correlation is weak. Furthermore, a determination is made that the solid object is present in the superposition area if the image feature values in each image have locationally overlapping portions. In such a circumstance, the image is composited with the blend rate of the image with a greater image feature value set large. |
