Method and apparatus for high speed image acquisition

• US 6,025,854 A
• Filed: 12/31/1997
• Issued: 02/15/2000
• Est. Priority Date: 12/31/1997
• Status: Expired due to Term

• Current Assignee: Cognex Corporation, Russ Weinzimmer
• Original Assignee: Cognex Corporation
• Inventors: Olson, Steven J., Anderson, Kurt M., Hinz, Robert C.
• Primary Examiner(s): Tung, Kee M.
• Application Number:
  US09/002,264
• Time in Patent Office: 776 Days
• Field of Search: 345/511, 345/508, 345/513, 345/501, 345/507, 345/509, 345/515, 345/517, 345/521
• US Class Current: 345/538
• CPC Class Codes: G06T 1/60 Memory management
• CPC Class Codes: H04N 5/4448 for frame-grabbing
• CPC Class Codes: H04N 7/012 Conversion between an interlaced and a progressive signal for graphics images G09G2310/0229

Abstract

The invention provides methods and systems for forming and image from a stream of non-interleaved image date, e.g., a stream of non-contiguous rows of pixels. Such a system includes a memory, in which the image is to be formed (and stored), and a controller, for initiating memory transfers that move rows of pixels from the stream to their (the rows'"'"') respective locations in memory. The controller initiates those transfers such that (i) the rows of pixels are transferred to the memory in the order in which they are received from the stream, and (ii) each row of pixels is transferred directly to the location in memory corresponding to that row'"'"'s respective position in the image.

Claims

1. A system for forming an image from a stream of non-interleaved rows of pixels, the system comprising

A. a host device that includes a host memory having a plurality of addressable storage locations in which the image is to be stored, the host device specifying respective locations of the host memory into which successive rows received from the stream are to be stored,

B. image acquisition logic that is coupled to the host device, the image acquisition logic including a controller for at least initiating a plurality of memory transfers, each memory transfer for storing a respective row of pixels in a respective location in the memory specified by the host device, and

C. the controller initiating the transfers such that

(i) the rows of pixels are transferred to the host memory substantially in an order in which they are received from the stream, and

(ii) the rows of pixels are transferred substantially directly to locations in the host memory corresponding to relative positions of those rows in the image.

2. A system according to claim 1 wherein the controllers transfers rows of pixels to respective locations in the host memory by direct memory access (DMA) transfers.

3. A system according to claim 2, wherein the controller initiates each memory transfer by issuing an instruction to transfer a row of pixels to a specified location in the memory.

4. A system according to claim 3, wherein the host comprises a table specifying respective locations in the host memory to which successive rows received from the stream are to be stored.

5. A system according to claim 1, in which the stream comprises successive fields, each including a plurality of respective rows of pixels of the image.

6. A system according to claim 5, in which the stream comprises two fields, a first field including rows of pixels for odd rows in the image, a second field including rows of pixels for even rows in the image.

7. A system according to claim 6, wherein the controller responds to receipt of the first field for initiating memory transfers for transferring the rows of pixels therein to locations in the host memory corresponding to odd rows in the image.

8. A system according to any of claims 6 and 7, wherein the controller responds to receipt of the second field in the stream for initiating memory transfers for transferring the rows of pixels therein to locations in the host memory corresponding to even rows in the image.

9. In an image acquisition system, the improvement for forming an image received as a stream of non-interleaved rows of pixels, comprising

A. a host device comprising a first memory having a plurality of addressable storage locations in which the image is to be stored, the host device specifying respective locations of the host memory into which successive rows received from the stream are to be stored,

B. image acquisition logic coupled to the host device, the image acquisition logic including a second memory for storing pixels substantially in an order they are received from the stream,

C. a bus coupled to the first and second memories for carrying pixels therebetween,

D. the image acquisition including a controller for at least initiating a plurality of memory transfers over the bus, each transfer moving a pixels from the second memory to a respective location in the first memory specified by the host device, such that

(i) the pixels are transferred from the second memory to the first memory substantially in an order in which the pixels are received from the stream of non-interleaved rows of pixels, and

(ii) the pixels are transferred substantially directly to locations in the first memory corresponding to their relative positions in the image.

10. A system according to claim 9, wherein the pixels are transferred from the second memory to the first memory by direct memory access (DMA) transfer.

11. A system according to claim 10, wherein the controller initiates each memory transfer by issuing an instruction to transfer pixels from the second memory to a specified location in the first memory.

12. A system according to claim 11, wherein the controller issues DMA instructions for the transfer of a row of pixels at a time from the second memory to the first memory.

13. A system according to claim 12, comprising a table of pointers specifying respective locations in the first memory to which successive rows of are to be transferred from the second memory.

14. A system according to claim 13, wherein the image acquisition logic comprises a third memory, coupled to the controller, for storing at least one of the pointers from the table.

15. A system according to claim 14, wherein the pointers are transferred from the table to the third memory by DMA transfer.

16. A frame grabber for storing to an addressable memory in a host device an image received by the frame grabber as a stream of non-interleaved rows of pixels, the frame grabber comprising:

A. a first FIFO memory for at least temporarily storing pixels as they are received from the stream,

B. a second FIFO memory for at least temporarily storing pointers specifying locations in the host device memory to which successive rows of pixels are to be transferred from the first FIFO memory,

C. a controller, coupled to the first and second FIFO memories, for at least initiating a plurality of memory transfers from the first FIFO memory to locations in the host device memory specified by successive pointers in the second FIFO memory,

D. the controller initiating the transfers such that;

(i) the pixels are transferred from the first FIFO memory to the host device memory in an order in which the pixels are received from the stream, and

(ii) the pixels are transferred substantially directly to locations in the host device memory corresponding to their respective positions in the image.

17. A frame grabber according to claim 16, wherein the pointers are loaded into the second FIFO memory by DMA transfers.

18. A frame grabber according to claim 17, wherein the controller initiates DMA transfers from a table of pointers in the host device to the second FIFO memory.

19. A frame grabber according to claim 16, in which the stream comprises successive fields, each including a plurality of respective rows of pixels making up the image.

20. A frame grabber according to claim 19, in which the stream comprises a first field including rows of pixels for odd rows in the image, and a second field including rows of pixels for even rows in the image.

21. A frame grabber according to claim 20, wherein the controller responds to receipt from the stream of the first field for initiating memory transfers to locations in the host device memory corresponding to odd rows in the image.

22. A frame grabber according to any of claims 20 and 21, wherein the controller responds to receipt from the stream of the second field for initiating transfers to locations in the host device memory corresponding to even rows in the image.

23. A frame grabber according to claim 22, comprising a third memory for storing a first pointer to a table in the host device containing pointers to locations in the host device memory corresponding to successive odd rows of the image and, thereafter, to successive even rows in the image, and a second pointer to a table in the host computer containing pointers to locations in the host computer memory corresponding to successive even rows of the image and, thereafter, to successive odd rows in the image.

24. A frame grabber according to claim 23, wherein the controller responds to receipt from the stream of the first field for initiating a DMA transfer with the first pointer, thereby causing the second memory to be loaded with pointers from the first table.

25. A frame grabber according to claim 23, wherein the controller responds to receipt from the stream of the second field for initiating a DMA transfer with the second pointer and, thereby, to cause the second memory to be loaded with pointers from the second table.

26. A method for transferring an image from an image acquisition device to a host memory of a host device to which the image acquisition device is coupled, the image being represented by a stream of non-interleaved rows of pixels received by the image acquisition device, the method comprising the steps of

specifying with the host device respective locations in the host memory into which successive rows received from the stream are to be stored,

initiating with the image acquisition device a plurality of memory transfers, each memory transfer storing a respective row of pixels from the stream in a respective location specified by the host device in the host memory, such that

(i) the rows of pixels are transferred substantially in an order in which they are received from the stream of non-interleaved rows of pixels, and

(ii) the rows of pixels are transferred substantially directly to locations in host memory corresponding to their relative positions in the image.

27. A method according to claim 26, comprising transferring the rows of pixels from the stream to their respective locations in the host memory by direct memory access (DMA) transfer.

28. A method according to claim 27, comprising:

A. loading a table in the host device with successive pointers to locations in the host memory to which successive respective rows received from the stream are to be transferred,

B. wherein the initiating step comprising initiating each DMA transfer by issuing a DMA instruction to transfer a row of pixels to a location specified by a next successive pointer in the table.

29. A method according to claim 26, in which the stream comprises successive fields, each including a plurality of respective rows of pixels of the images.

30. A method according to claim 29, wherein the initiating step comprises

A. responding to receipt of a first field in the stream for initiating memory transfers to locations in the host memory beginning with a first set of rows in the images,

B. responding to receipt of a second field in the stream from initiating memory transfers to locations in the host memory beginning with a second set of rows in the images, and

C. where the first and second sets are substantially exclusive of one another.

31. A method according to claim 30, wherein the first set of rows in the image are odd-numbered rows and wherein the second set of rows in the image are even-numbered rows.

32. An image acquisition method for transferring an image received by an image acquisition device as a stream of non-interleaved rows of pixels to a host device to which the image acquisition device is coupled, the method comprising:

A. storing in a first memory in the image acquisition device pixels substantially in an order that they are received from the stream of non-interleaved rows of pixels,

B. specifying with the host device respective locations in the host memory into which successive rows received from the stream are to be stored,

C. initiating a plurality of memory transfers, each memory transfer for moving a respective row of pixels from the first memory to a respective location specified by the host device in a second memory in the host, such that

(i) the pixels are transferred from the first memory to the second memory substantially in an order that they are received from the stream of non-interleaved rows of pixels; and

(ii) the pixels are transferred substantially directly to locations in the second memory corresponding to their respective locations in the image.

33. A method according to claim 32, comprising the steps of transferring the pixels from the first memory to the second memory by direct memory access (DMA) transfer.

34. A method according to claim 33, wherein the initiating step comprises initiating each DMA transfer by issuing an instruction to transfer pixels to a location specified by a next successive pointer in the table.

35. A method according to claim 34 wherein the initiating step comprises issuing the DMA instructions for the transfer of one row of the pixels at a time from the first memory to the second memory.

36. A method of operating a frame grabber for storing an image, received as a stream of non-interleaved rows of pixels, to a host computer, the method comprising the steps of:

A. storing a first FIFO memory pixels substantially in an order which they are received from the stream,

B. storing to a second FIFO memory pointers specifying locations in the host computer memory to which successive rows of pixels are to be transferred from the first FIFO memory,

C. initiating a plurality of memory transfers from the first FIFO memory to locations in the host computer memory specified by successive pointers in the second FIFO memory, such that

(i) the pixels are transferred from the first FIFO memory to the host computer memory substantially in an order in which they are received from the stream,

(ii) the pixels are transferred substantially directly to locations in the host computer memory corresponding to their respective positions in the image.

37. A method according to claim 36, wherein the pointers are loaded into the second FIFO memory by DMA transfers.

38. A method according to claim 37, in which the stream comprises successive fields, each including a plurality of respective rows of pixels of the image.

39. A method according to claim 38, wherein the initiating step comprises

A. responding to receipt of a first field in the stream for initiating memory transfers to locations in the host computer memory corresponding to a first set of rows in the image,

B. responding to receipt of a second field in the stream for initiating memory transfers to locations in the host computer memory to a second set of rows in the image, and

C. wherein the first and second sets are substantially exclusive of one another.

40. A method according to claim 39, wherein the first set of rows in the image are odd rows and wherein the second set of rows in the image are even rows.