| abstract |
The invention discloses a non-contact object material and shape identification method, a storage medium and a computing device. The object to be tested is placed in a designated detection area of a sensing plane, and the ID and RSS of each tag in the process of communication between an RFID reader and a tag are collected. value and phase parameters, identify the shape and material of the object to be tested; perform similarity calculation, and determine the classification of the object to be tested according to the obtained results as seen items and unseen items; if it is a seen item, it will collect The obtained RSS and phase signals are extracted to obtain the optimal features, and the material and shape of the item to be tested are identified; if it is an unseen item, the text feature that is closest to the RF feature similarity of the item to be tested is searched in the hidden space. The corresponding text with the highest text feature similarity is used as the recognition category of the item to be tested, and the material and shape of the item to be tested are identified. The invention has low cost, realizes cross-environment, determines the shape and material of the item at the same time, and recognizes the unseen item, that is, the item that does not exist in the training set. |