Depth-spatial frequency-response assessment

1. A method to calibrate a depth-imaging camera to improve fidelity to depth-change abruptness of an imaged subject, the method the comprising:

digitally generating, with the depth-imaging camera, a machine-readable calibration depth image of a calibration subject arranged in a field of view of the depth-imaging camera;

machine processing the calibration depth image in a spatial domain to obtain a machine-readable measure of the fidelity in the spatial domain;

machine processing the measure of the fidelity in the spatial domain to obtain a machine-readable measure of the fidelity in a frequency domain;

extracting a deconvolution parameter value from the measure of the fidelity in the frequency domain;

digitally generating, with the depth-imaging camera, an in-use depth image of a user-defined subject; and

in logic of the depth-imaging camera, deconvoluting the in-use depth image based on the deconvolution parameter value.

2. The method of claim 1 further comprising outputting, from the depth-imaging camera, a deconvoluted machine-readable depth image.

3. The method of claim 1 wherein the measure of the fidelity in the spatial domain includes an edge-spread function of the calibration depth image.

4. The method of claim 3 wherein processing the measure of the fidelity in the spatial domain includes differentiating the edge-spread function to obtain a line-spread function of the calibration depth image.

5. The method of claim 4 wherein processing the measure of the fidelity in the spatial domain includes computing a Fourier transform of the line-spread function to obtain a depth-spatial frequency response (DSFR) function.

6. The method of claim 5 wherein deconvoluting the in-use depth image includes direct inverse filtering and/or Wiener filtering in an amount to compensate for the DSFR function.

7. The method of claim 1 wherein deconvoluting the in-use depth image includes increasing the depth-change abruptness of a locus of the in-use depth image.

8. The method of claim 1 further comprising denoising the in-use depth image in logic of the depth-imaging camera prior to deconvoluting the in-use depth image.

9. The method of claim 8 wherein said processing and extracting are repeated for each of a plurality of denoising-filter parameter values, yielding a corresponding plurality of deconvolution parameter values, the method further comprising selecting from among the deconvolution parameter values based on an amount of denoising applied to the in-use depth image.

10. The method of claim 1 wherein said processing and extracting are repeated for each of a plurality of temperatures, yielding a corresponding plurality of deconvolution parameter values, the method further comprising selecting from among the deconvolution parameter values based on temperature measured during acquisition of the in-use depth image.

11. A depth-imaging camera calibrated to improve fidelity to depth-change abruptness of an imaged subject, the depth-imaging camera comprising:

an array of light-sensing pixel elements;

electronic memory;

logic configured to address each of the pixel elements, and, for each pixel element addressed, to compute depth to a subject locus imaged on that pixel element, and to generate an in-use depth image including the computed depth for each pixel element; and

logic configured to deconvolute the in-use depth image based on a deconvolution parameter stored in the electronic memory, the deconvolution parameter extracted from a machine-readable measure of the fidelity to the depth-change abruptness in a frequency domain obtained by processing a machine-readable measure of the fidelity in a spatial domain, the measure of the fidelity in the spatial domain obtained by processing a calibration depth image in the spatial domain.

12. The depth-imaging camera of claim 11 wherein the computed depth is a composite based on addressing each pixel element two or more times, and wherein the logic configured to deconvolute the in-use depth image applies deconvolution for each of the two or more times that the pixel element is addressed.

13. The depth-imaging camera of claim 12 further comprising a modulated IR emitter, wherein addressing each pixel element two or more times includes addressing at two or more different modulation frequencies of the IR emitter.

14. The depth-imaging camera of claim 12 wherein the computed depth is a composite based on addressing each pixel element two or more times and wherein the logic configured to deconvolute the in-use depth image applies deconvolution to the composite.

15. The depth-imaging camera of claim 11 further comprising a temperature sensor,

wherein a different deconvolution parameter value is stored in the electronic memory for each of a plurality of temperatures, and wherein the logic configured to deconvolute the in-use depth image is further configured to select the deconvolution parameter value based on a reading of the temperature sensor.