An approach to estimate object depth from a visual projection scene for eye-in-hand robot applications

Abstract EN

In this paper, the depth of a monitored object using eye-in-hand configuration is Estimated and verified by application of a proposed approach that is based on a pre-calibration process of the image and the physical area of the monitored object with respect to camera height.

In addition, neural network approach is also used to manipulate the acquired vision data in order to obtain the required object depth.

Both proposed approaches implemented using the eye-in-hand configuration and showed good results to improve the overall capability of the eye-in-hand configuration to perform intelligent robot tasks.