Image processing apparatus, imaging apparatus, and image processing method
A field of view of a captured image of a lensless camera can be controlled, and a configuration of generating a restored image including part of an imaging region is realized. Included is a signal processing unit that receives observed image signals as output of an image sensor of a lensless camera to generate a restored image of a restored image region including part of a captured image region of the lensless camera. The signal processing unit includes: a restored image region corresponding mask matrix calculation unit that calculates a restored image region corresponding mask matrix applied to generate the restored image; and an image estimation unit that subtracts observed image signals outside of the restored image region not included in the restored image region from the observation image signals to calculate observed image signals inside of the restored region and that executes arithmetic processing of the observed image signals inside of the restored region and a pseudo-inverse matrix or an inverse matrix of the restored image region corresponding mask matrix to generate the restored image.
1. An image processing apparatus, comprising:
a signal processing unit that receives observed image signals as output of an image sensor of a lensless camera to generate a restored image of a restored image region including part of a captured image region of the lensless camera, wherein
the signal processing unit generates the restored image by executing the arithmetic processing of observed image signals in a restored region generated by subtracting observed image signals not included in the restored image region from the observed image signals and executing the arithmetic processing of the observed image signals in the restored region and an inverse matrix or a pseudo-inverse matrix of an applied matrix to generate the restored image.
2. The image processing apparatus according to claim 1 , wherein
the signal processing unit includes a restored image region corresponding matrix calculation unit that generates a matrix applied to generate the restored image based on correspondence information of emitted light from sampling points in the restored image region and observed values of the image sensor.
3. The image processing apparatus according to claim 1 , wherein
the signal processing unit includes a restored image region corresponding matrix calculation unit that generates a matrix applied to generate the restored image of P×N defined by the number of sampling points (P) in the restored image region and the number of pixels (N) of the lensless camera.
4. The image processing apparatus according to claim 1 , wherein
the signal processing unit includes an image estimation unit that generates the restored image by executing the arithmetic processing of observed image signals in the restored region, the observed image signals being calculated by subtracting the observed image signals not included in the restored image region from emitted light estimation values of sampling points in the captured image region calculated based on an inverse matrix of a matrix applied to generate a restored image of an entire imaging range of the lensless camera and based on the observed image signals, and an inverse matrix of the matrix applied to generate the restored image of the entire imaging range of the lensless camera.
5. The image processing apparatus according to claim 1 , wherein
the signal processing unit further includes a captured image region corresponding mask matrix calculation unit that generates a matrix applied to the captured image region based on correspondence information of emitted light from sampling points in the captured image region and observed values of the image sensor.
6. The image processing apparatus according to claim 1 , wherein
the signal processing unit sets density of sampling points in the restored image region to density higher than density of sampling points of other regions of a captured image to generate the restored image at high resolution.
7. The image processing apparatus according to claim 1 , wherein
the signal processing unit partitions wavelength of emitted light of sampling points in the restored image region to generate a restored image in a plurality of different wavelength light units.
8. The image processing apparatus according to claim 1 , wherein
the signal processing unit partitions polarized components of emitted light of sampling points in the restored image region to generate a restored image in a plurality of different polarized component units.
9. An imaging apparatus, comprising:
a mask in which transmission regions and non-transmission regions of light are set as a two-dimensional pattern;
an imaging unit including an image sensor that receives the light through the mask; and
a signal processing unit that receives observed image signals as output of the image sensor to generate a restored image of a restored image region including part of a captured image region of the imaging unit, wherein
the signal processing unit generates the restored image by executing arithmetic processing of observed image signals in a restored region generated by subtracting observed image signals not included in the restored image region from the observed image signals and executing the arithmetic processing of the observed image signals in the restored region and an inverse matrix or a pseudo-inverse matrix of an applied matrix to generate the restored image.
10. An image processing method executed by an image processing apparatus, wherein
the image processing apparatus includes a signal processing unit that receives observed image signals as output of an image sensor of a lensless camera to generate a restored image of a restored image region including part of a captured image region of the lensless camera, and
the signal processing unit generates the restored image by executing arithmetic processing of observed image signals in a restored region generated by subtracting observed image signals not included in the restored image region from the observed image signals and executing the arithmetic processing of the observed image signals in the restored region and an inverse matrix or a pseudo-inverse matrix of an applied matrix to generate the restored image.