Optical ranging apparatus

• US 5,076,687 A
• Filed: 08/28/1990
• Issued: 12/31/1991
• Est. Priority Date: 08/28/1990
• Status: Expired due to Fees

• Current Assignee: Massachusetts Institute of Technology
• Original Assignee: Massachusetts Institute of Technology
• Inventors: Adelson, Edward H.
• Primary Examiner(s): Buczinski, Stephen C.
• Application Number:
  US07/574,417
• Time in Patent Office: 490 Days
• Field of Search: 356/4, 250/201.7, 250/201.8, 354/402
• US Class Current: 356/4.04
• CPC Class Codes: G01C 3/08 Use of electric radiation detectors

Abstract

An optical ranging apparatus is provided which resolves the depth of objects in an object field of a main lens. Light is directed by the main lens to a lenticular array consisting of an array of lenticules each of which generates an image of the lens surface. Each of the generated images is directed to a photodetector array divided into macropixels each of which receives one of the lenticule images. Each macropixel is made up of a number of subpixels, each of which is a discrete photodetector element and generates an electrical signal proportional to the intensity of light upon it. A data processor receives the signals from the subpixels and assembles subimages, each of which consists of one subpixel from each macropixel. Each subimage represents a view of the entire object field through one portion of the main lens. The data processor determines depth of objects in the object field by comparing the parallax between subimages. A diffuser may be provided between the object field and the lenticular array to low pass filter light from the object field and reduce aliasing. A field lens may also be provided between the main lens and the lenticular array to redirect light from the main lens and prevent skewing of the light reaching the macropixels. The images from the lenticular array may be relayed to a photodetector array set back from the lenticular array by using a relay lens.

Claims

1. An optical depth resolver, comprising:

a converging lens receiving light from objects in an object field and directing the received light to an image field of the converging lens;

a plurality of imaging elements distributed about the image field of the converging lens in the image field of the lens such that each imaging element receives light from the converging lens and forms an image from the light it receives;

a photodetector array divided into a plurality of macropixels each of which is made up of a plurality of subpixels, each macropixel receiving an image formed by one of the imaging elements such that each subpixel of a macropixel receives light from one portion of the lens as imaged by one of the imaging elements, each subpixel generating an electrical signal indicative of the intensity of light incident upon it; and

a data processor receiving the electrical signals from the subpixels and processing the signals from different subpixels, the data processor identifying subimages of subpixels having like spatial positions within different macropixels and performing, across the image field, a plurality of local displacement estimates of image displacement between subimages, to create a range image of depths of objects across the object field.

2. An optical depth resolver according to claim 1 wherein the plurality of imaging elements comprises a lenticular array.

3. An optical depth resolver according to claim 2 wherein the lenticular array comprises an array of cylindrical lenses.

4. An optical depth resolver according to claim 2 wherein the lenticular array comprises an array of spherical lenses.

5. An optical depth resolver according to claim 4 wherein the spherical lenses are arrayed in a rectangular lattice.

6. An optical depth resolver according to claim 1 wherein each of the imaging elements receives light from the entire surface of the converging lens.

7. An optical depth resolver according to claim 1 wherein the photodetector array is a CCD array.

8. An optical depth resolve according to claim 1 wherein each subimage is representative of light passing through a different region of the converging lens.

9. An optical depth resolver according to claim 8 wherein the data processor assembles the subimages and compares the parallax from one subimage to another in determining depth of objects in the object field.

10. An optical depth resolver according to claim 8 wherein the shift of intensity peaks from one subimage to another is detected by the data processor and used determining depth of objects in the object field.

11. An optical depth resolver according to claim 1 wherein there is one macropixel for each image generated by an imaging element.

12. An optical depth resolver according to claim 1 wherein the data processor runs a least squares processing method in making determinations of depth of objects in the object field.

13. An optical depth resolver according to claim 1 further comprising a weak diffuser which diffuses the light from objects in the object field prior to the light being received by the imaging elements.

14. An optical depth resolver according to claim 13 wherein the diffuser is a piece of glass with random surface variations.

15. An optical depth resolver according to claim 1 further comprising a field lens positioned between the converging lens and the imaging elements, and redirecting the light from the converging lens such that the light from the converging lens appears to have been focused from infinity.

16. An optical depth resolver according to claim 1 further comprising a relay lens between the imaging elements and the photodetector array, the relay lens redirecting the images from the imaging elements to the photodetector array.

17. An optical depth resolver according to claim 16 wherein the relay lens refocuses the images generated by the imaging elements.

18. An optical depth resolver according to claim 17 wherein the distribution of the images from the imaging elements across the surface of the photodetector array is controlled by the focusing of the relay lens.

19. An optical depth resolver according to claim 16 wherein the relay lens is part of a television camera.

20. An optical depth resolver as claimed in claim 1 wherein the data processor estimates are based on derivatives of local intensity between subimages for patches of macropixels.

21. An optical depth resolver as claimed in claim 20 wherein the estimates are based on derivatives in two dimensions.

22. An optical depth resolver as claimed in claim 1 wherein the data processor applies a weighting mask to the electrical signals generated by the photodetector array in order to adjust the signals for misalignment of the imaging elements with the photodetector array.

23. An optical depth resolver as claimed in claim 1 wherein the data processor processes subpixels of at least three subimages in performing each local displacement estimate.

24. An optical depth resolver comprising:

a converging lens receiving light from objects in an object field and directing the received light to an image field of the converging lens;

a lenticular array having a two dimensional array of lenticules distributed about the image field of the converging lens such that each lenticule receives light directed from a different portion of the object field and receives the directed light at a different angle relative to an optic axis of the main lens, each lenticule forming an image from the light it receives;

a photodetector array divided into a plurality of macropixels each of which is made up of a plurality of subpixels, each macropixel receiving the image formed by one of the lenticules such that each subpixel of a macropixel receives light from one portion of the converging lens as imaged by one lenticule, each subpixel generating an electrical signal indicative of the intensity of light incident upon it;

a data processor receiving the electrical signals from the subpixels and processing the subpixel signals to detect variations in intensity of light directed from a plurality of locations in the object field relative to the portion of the converging lens through which the light passes, the data processor identifying subimages of subpixels having like spatial positions within different macropixels and performing, across the image field, estimates of depths of objects across the object field based on local derivatives of intensity between subimages to create a range image.

25. An optical depth resolver according to claim 24 wherein the photodetector array is a CCD array.

26. An optical depth resolver according to claim 24 further comprising a weak diffuser diffusing light between the object field and the lenticular array.

27. An optical depth resolver according to claim 24 further comprising a field lens between the converging lens and the lenticular array, the field lens redirecting the light from the converging lens such that it appears to have been focused from infinity.

28. An optical depth resolver according to claim 24 further comprising a relay lens between the lenticular array and the photodetector array, the relay lens redirecting the images from the lenticular array to the photodetector array.

29. An optical depth resolver as claimed in claim 24 wherein the estimates are based on derivatives in two dimensions.

30. An optical depth resolver as claimed in claim 24 wherein the data processor applies a weighting mask to the electrical signals degenerated by the photodetector array in order to adjust the signals for misalignment of the lenticular array with the photodetector array.

31. An optical depth resolver as claimed in claim 24 wherein the data processor processes at least three subimages to perform each estimate.

32. A method of making depth measurements, comprising:

providing a converging lens receiving light from objects in an object field and directing the received light to an image field of the converging lens;

receiving said light from the converging lens with a plurality of imaging elements distributed about the image field of the converging lens, each imaging element forming an image from light it receives;

receiving the images formed by the imaging elements with a photodetector array divided into a plurality of macropixels each of which is made up of a plurality of subpixels, each macropixel receiving an image formed by one of the imaging elements such that each subpixel of a macropixel receives light from one portion of the lens as imaged by one of the imaging elements, each subpixel generating an electrical signal indicative of the intensity of light incident upon it; and

receiving the electrical signals from the subpixels with a data processor and processing the signals from different subpixels to create a range image of estimates of depths of objects across the object field.

33. A method according to claim 32 wherein receiving said light from the converging lens with a plurality of imaging elements comprises receiving said light from the converging lens with a lenticular array.

34. A method according to claim 32 wherein receiving said light from the converging lens with a plurality of imaging elements comprises receiving said light with the imaging elements such that each imaging element receives light from the entire surface of the converging lens.

35. A method according to claim 32 wherein receiving the images formed by the imaging elements further comprises arranging the subpixels of each macropixel of the photodetector array in a similar positional distribution such that a group of subpixels, each from a similar position in a different macropixel, together form a subimage representative of light passing through one region of the converging lens.

36. A method according to claim 35 wherein processing the electrical signals from different subpixels further comprises assembling said subimages and comparing the parallax from one subimage to another in determining depth of objects in the object field.

37. A method according to claim 35 wherein processing the electrical signals from different subpixels further comprises detecting the shift of intensity peaks from one subimage to another in determining depth of objects in the object field.

38. A method according to claim 32 further comprising diffusing said light from objects in the object field with a weak diffuser.

39. A method according to claim 32 further comprising redirecting the light from the converging lens to the imaging elements with a field lens.

40. A method according to claim 32 further comprising redirecting the images from the imaging elements to the photodetector array with a relay lens.

41. A method as claimed in claim 32 wherein the data processor estimates are based on derivatives of local intensity between subimages for patches of macropixels.

42. A method as claimed in claim 41 wherein the estimates are based on derivatives in two dimensions.

43. A method as claimed in claim 32 wherein the data processor applies a weighting mask to the electrical signals generated by the photodetector array in order to adjust the signals for misalignment of the lenticular array with the photodetector array.

44. A method as claimed in claim 32 wherein the data processor processes at least three subimages to determine each depth of an object.