We used the LabelBox online toolbox to create this data set. It consists of 6 different cluttered environments: a laboratory 3 different industrial-like environments a corridor hall We proposed to split the data set into three sets - training, validation, and test sets - as follows: (1) the training set has 156 frames equally distributed by the laboratory and 1 workshop; (2) the validation set is also divided into two environments - the corridor (158 frames) and a different workshop (66 frames); (3) the test set consists of 145 frames collected on a neat hall with overshadowed and over-lightened landmarks in different planes; a classroom laboratory with various electronic equipment arranged in an orderly manner; and a very challenging scenario with multiple pieces of machinery spread out all over the place. One should filter out images with no markers.

This work was supported by Project PRODUTECH II SIFPOCI-01-0247-FEDER-024541, and partially supported by Project SeaAI-FA 02 2017 011.