System and method for object recognition

1. A computer implemented method of object recognition of an object to be identified, the method comprising the steps of:

acquiring, by a mobile device, a plurality of pictures of said object;

sending the acquired pictures to a cloud server;

reconstructing, by the cloud server, a 3D model of the object, to derive the 3D points cloud reconstruction of the object, wherein the step of reconstructing, by the cloud server, a 3D model of the object comprises:

obtaining a plurality of pictures of the object;

extracting a plurality of key points that can be correlated in said plurality of pictures of the object, wherein at least two pictures of the plurality of pictures show at least two different viewpoints of the object;

placing the key points on the object;

defining a plurality of 3D slices of the object, wherein a 3D slice comprises at least one key point; and

adding the 3D slices to build a reconstructed 3D points cloud of the object; and

performing a 3D match search in a 3D database using the 3D points cloud reconstruction, to identify the object, the 3D match search comprising comparing the 3D points cloud reconstruction of the object with 3D points clouds of known objects stored in the 3D database; and

wherein the step of acquiring a plurality of pictures comprises:

extracting said pictures from a video sequence, the method comprising dynamically adjusting the acquisition parameter depending on the 3D points cloud reconstruction, wherein pictures from the video sequence are saved every «n» frame, and wherein «n» being adjusted dynamically or by the user, in particular wherein n is given a higher value at the start of the method and decreases as the reconstruction becomes more accurate.

2. The method according to claim 1 , further comprising:

performing a first search match in a first database, the first search match being performed using 2D recognition techniques including at least one of Optical Character Recognition, Scale Invariant Feature Transform based imaging, color gradient analysis.

3. The method of claim 1 , wherein the steps of the method are performed concurrently until the object has been identified in at least one of the first database and the 3D database, and particular pictures are acquired as long as the object has not been acquired or until a time out has been reached.

4. The method of claim 1 , wherein the search match in the 3D database includes at least one of machine learning or 3D geometric comparison.

5. The method of claim 1 , further comprising:

displaying in real time information pertaining the method on the mobile device, wherein the mobile device comprises an input device allowing a user to enter input data concerning the object to be identified, and wherein the 2D or 3D match search are adapted depending on the input data.

6. A computer program product comprising a non-transitory computer medium having stored therein to cause a computer to perform the steps of a method according to claim 1 .

7. The method according to claim 1 , wherein the step of reconstructing, by the cloud server, a 3D model of the object comprises:

defining an initial set of searchable points in a first picture of the plurality of pictures; and

identifying some of the searchable points in the remaining pictures of the plurality of pictures, to extract the key points.

8. The method of claim 1 , the step of reconstructing, by the cloud server, a 3D model of the object further comprising:

computing a calibration matrix in a reference frame to derive a relative measurement system.

9. The method of claim 8 , wherein the 3D slices are added in the obtained relative measurement system.