深度图像的边缘检测

Borderline Inspection of Depth-Image

用强度图像提取景物边界的主要缺点是当景物间灰度差较小以及噪声存在时，很难得到精确的边界．本文通过单条纹结构光获取景物的距离图像，用距离图像对物体进行边缘提取，可以弥补强度图像中由于景物间灰度差较小而检测不到的边界，减少由于噪声产生的错误边界，从而得到更可靠、准确的物体边界，为物体的特征提取提供依据．

It is very difficult to obtain the accurate borderline of object When distilling the borderline of object with less gray level from its intensity image with disturbance. In this paper, the depth - image of object is obtained using the sigle -stripe structured - light, and the borderline of object is distilled from its depth-image, which can fetch up the borderline which is not inspected because of the less gray level of the object and reduce error borderline produced by disturbance. So more reliable and accurate borderlines can be obtained and the base of distilling the characters of object is set up.