Device and method of object detection

A device comprising:

• a single focusing optics;
• a 2D imager;
• a 3D time-of-flight imager;
• and
• a dichroic mirror positioned between the optics and the two imagers such that both imagers see the same scene as focused by the single, shared focusing optics;
• circuitry adapted to accept image pairs, from a single scene at a single point in time, a first 2D image digital dataset and a corresponding first 3D time-of-flight image digital dataset, from the data storage and processing modules, and then use information in the first 3D time-of-flight image digital dataset to enhance the first 2D image digital dataset, and adapted to output, store or further process the now enhanced first 2D image digital dataset (e-RGB).

A method of enhancing two-dimensional images comprising the steps:

• (a) acquiring a pair of images at the same time, of the same scene, using a single focusing optical path, wherein a 3D dataset is acquired with a time-of-flight sensor and a 2D image is acquired by 2D image sensor, and the 3D dataset and the 2D image make up the pair of images;
• and
• (b) using distance data, amplitude data, or both, from the 3D dataset to enhance selected portions of the 2D image by replacing or altering groups of pixels in the 2D image, generating an e-RGB image;
• wherein the single focusing optical path comprises a dichroic mirror positioned between the optics and the two imagers such that both the 2D image sensor and the time-of-flight sensor see the same scene as focused by the single, shared focusing optics.

A method of generating bounding boxes for object identification in an images comprising the steps:

• (a) acquiring a pair of images at the same time, of the same scene, using a single focusing optical path, wherein a 3D dataset is acquired with a time-of-flight sensor and a 2D image is acquired by 2D image sensor, and the 3D dataset and the 2D image make up the pair of images;
• (b) using distance data, amplitude data, or both, from the 3D dataset to identify a region, groups of pixels, or objects wherein the distance of at least a majority of pixels in the region are at the same distance, within a distance bound, to create a first bounding box;
• and
• (c) transferring the bounding box to the 2D image for the purpose of object identification in the 2D image;
• wherein the optical path comprises a dichroic mirror positioned between the optics and the two imagers such that both the 2D image sensor and the time-of-flight sensor see the same scene as focused by the single, shared focusing optics.