Devices and methods for high dynamic range video

1. A method for producing real-time HDR video, the method comprising:

simultaneously capturing, with multiple image sensors, a plurality of images that differ by light level;

streaming pixel values from each of the multiple image sensors in a frame independent-manner through a pipeline on a processing device, wherein the pipeline includes

a sync module that synchronizes the pixel values of the images as the pixel values stream through the processing device,

a kernel operation that identifies saturated pixel values among the synchronized pixel values; and

merging the pixel values to produce an HDR image.

2. The method of claim 1 , wherein the merging step excludes at least some of the saturated pixel values from the HDR image.

3. The method of claim 2 , wherein the sync module comprises line buffers that circulate early-arriving pixel values and then release the early-arriving pixel values simultaneously with corresponding, later-arriving pixel values.

4. The method of claim 3 , wherein the multiple image sensors all capture images simultaneously through a single lens, the method further comprising receiving incoming light through the lens and splitting the light via at least one beamsplitter onto the multiple image sensors, wherein at least 95% of the light gathered by the imaging lens is captured by the multiple image sensors.

5. The method of claim 4 , wherein the multiple image sensors include at least a high exposure (HE) sensor and a middle exposure (ME) sensor, and wherein merging the sequences includes using HE pixel values that are not saturated and ME pixel values corresponding to the saturated pixel values.

6. The method of claim 5 , wherein the multiple sensors further include a low exposure (LE) sensor, and the method includes identifying saturated pixel values originating from both the HE sensor and the ME sensor.

7. The method of claim 5 , wherein streaming the pixel values through the kernel operation includes examining values from a neighborhood of pixels surrounding a first pixel on the HE sensor, finding saturated values in the neighborhood of pixels, and using information from a corresponding neighborhood on the ME sensor to estimate a value for the first pixel.

8. The method of claim 7 , wherein at least some of the saturated pixel values are identified before receiving values from all pixels of the multiple image sensors at the processing device.

9. The method of claim 8 , further comprising beginning to merge portions of the sequences while still streaming later-arriving pixel values through the kernel operation.

10. The method of claim 8 , wherein sequences of pixel values are streamed through the processing device and merged without waiting to receive pixel values from all pixels on the image sensors.

11. The method of claim 10 , wherein the processing device comprises one selected from the group consisting of a field-programmable gate array and an application-specific integrated circuit, and further wherein each of the image sensors comprises a color filter array, and the method further comprises demosaicing the HDR imaging after the merging.

12. The method of claim 11 , wherein multiple image sensors capture images that are optically identical except for light level.

13. The method of claim 3 , wherein streaming the pixel values through the kernel operation includes streaming the synchronized pixel values through a path within the processing device that momentarily places a value from the first pixel proximal to each value originating from the neighborhood of pixels.

14. The method of claim 11 , wherein the obtaining, streaming, and merging steps are performed by streaming the sequences of pixel values through a pipeline on the processing device such that no location on the processing device stores a complete image.

15. The method of claim 5 , wherein a first pixel value from a first pixel on one of the image sensors is identified as saturated if it is at least 90% of a maximum possible pixel value.