The autonomous robot object recognition algorithm based on multi-feature fusion

In a complex environment, the autonomous robot object tracking which is sole based on color feature is difficult to distinguish the object accurately, thus it causes the loss of the tracking object. An autonomous robot object recognition algorithm based on multi-feature fusion is presented. Firstly, the image is processed by the HSI dual adaptive threshold division method, and the run-length coding is used for regional integration on the target pixel, thus it can obtain the general scope of the target. Secondly, it can withdraw the outline of the object by using the Sobel border detecting method in the scope, and then these objects which have the same color as the target are removed through the outline contrast. The experimental result indicates that to a certain extent the method can effectively prevent the target from losing in a complex environment, and the method can enhance the object recognition accuracy and timeliness.