A processing device which obtains distance information of a subject, including: a calculation unit configured to calculate the distance information of the subject from a difference in blur degree of a plurality of images photographed by an imaging optical system; a correcting unit configured to correct the distance information using correction data in accordance with an image height in the imaging optical system; and an extraction unit configured to extract at least one frequency component from each of the plurality of images, wherein the calculation unit calculates the distance information from a difference in blur degree in the plurality of images in the at least one frequency component; and the correcting unit corrects the distance information using correction data in accordance with an image height in the at least one frequency component.

A passive imaging sensor includes a plurality of optical elements in which at least one includes one or more Micro-Electro-Mechanical System (MEMS) Micro-Mirror Arrays (MMAs) having a plurality of independently and continuously controllable mirrors that at least tip and tilt in 2 DOF and may tip, tilt and piston in 3 DOF. In an operational mode, the mirrors are tipped and tilted, and possibly pistoned, such that the optical radiation is focused at the pixelated detector to read out an image of the scene. NUC coefficients such as offset and/or gain are applied to either the output signals of the detector or to the image to form the NUC'd images. In a calibration mode, the mirrors are tipped and tilted and/or pistoned to spatially or temporally blur the image or to re-direct the FOV to one or more on-board calibration sources to generate a uniform image from which to calculate and update the NUC coefficients.

A phase difference AF processing unit (19) compares subject images between a region R1 and a region Rj (j=2 to m) in an AF area (53), and determines a phase difference detection pixel (52A, 52B) as a detection signal addition target with respect to a phase difference detection pixel (52A, 52B) in the region R1 among phase difference detection pixels (52A, 52B) in the region Rj. Further, the phase difference AF processing unit (19) adds up detection signals with respect to the phase difference detection pixels (52A, 52B) in the region R1 and the phase difference detection pixels (52A, 52B) which are addition targets, and generates a defocus amount (Df1) from a result of a correlation operation using detection signals after addition. A system control unit (11) drives a focus lens according to the defocus amount (Df1) to perform a focusing control.

A solid-state imaging device includes: a valid area including pixels that are not shielded from light; a first light-blocked area and a second light-blocked area each including pixels that are shielded from light; an analog-to-digital converting unit to convert the electric charge accumulated by the pixels belonging to the first light-blocked area, the valid area, and the second light-blocked area, to image data at a time; a signal reading unit to read light-blocked data obtained from the first light-blocked area and the second light-blocked area, and valid data obtained from the valid area, in units of pixels; a reference black level estimating unit to estimate a reference black level of the light-blocked data; and a level correction unit to correct, based on the estimated reference black level, a size of the valid data obtained simultaneously with the light-blocked data used in estimating the reference black level.

Various techniques are disclosed for providing object recognition using thermal imaging. Unique thermal features of an object such as a human face can be detected using a thermal imaging module. The thermal imaging module may be included in an authentication system that performs authentication operations for users of a secure system based on the detected thermal features. The thermal features may include a thermal map of a user's face. An object recognition system such as an authentication system may include a non-thermal imaging module that captures non-thermal images of the object. The object recognition system may recognize objects using thermal images and non-thermal images in separate object recognition operations or by combining the thermal and non-thermal images and performing object recognition operations using the combined image. A thermal imaging authentication system may help eliminate user passwords on phones, tablets, computers and/or other secure access systems.

Methods and apparatus for compensating for bad pixels in a sensor array. In embodiments, a detector system receives an image on a sensor array of pixels for a first frame via a lens when the lens and the sensor array are configured in a first positional relationship. The array includes at least one bad pixel. The system moves the lens and/or the sensor array based on a position of the at least one bad pixel in the image such that the lens and the sensor array are configured in a second positional relationship. The image on the sensor array for a second frame is received via the lens when the lens and the sensor array are configured in a second positional relationship. The system compensates for the location of the at least one bad pixel in the image for the first and second frames to output a processed image.

A method for updating a fixed pattern noise matrix comprises: calculating a first difference between a target and first different images in a video stream to obtain a first differential matrix; calculating a second difference between the target and second different images in the video stream to obtain a second differential matrix; identifying a set of candidate positions for fixed pattern noise by: locating first and second sets of positions in the first differential matrix at which a difference deviates from predetermined values, finding a set of overlapping positions between the first and second sets of positions, and adjusting the set of overlapping positions. The adjusted set of overlapping positions is used for fixed pattern noise. Furthermore, each position in the set of candidate positions is updated, wherein the updated fixed pattern noise value at each position is based on a value at a corresponding non-adjusted position in the differential matrix.

An image processing system includes a memory configured to store a plurality of reference images used for image quality tuning, an image signal processor configured to receive a plurality of captured images corresponding to the plurality of reference images and configured to generate a plurality of corrected images by being configured to perform a corresponding image processing operation among a plurality of image processing operations, and a tuning module configured to set parameters of the plurality of image processing operations based on the plurality of corrected images and the plurality of reference images.

Various techniques are disclosed to provide for detection of temporally anomalous flickering pixels. In one example, a method includes capturing, by a thermal imager of an imaging device, a plurality of thermal images in response to infrared radiation received from a uniform black body, wherein the thermal images comprise a plurality of pixels having associated pixel values. The method also includes determining, for each pixel, a standard deviation of the associated pixel values for the thermal images. The method also includes comparing the standard deviations with a threshold. The method also includes identifying a subset of the pixels as temporally anomalous pixels in response to the comparing. Additional methods, devices, and systems are also provided.

An image processing device and an image processing method are provided. The image processing device includes an image sensor configured to sense and output a raw image and an image signal processor configured to compare pixel values of the raw image with a reference level to classify a saturation color image and a non-saturation color image, calculate a first fixed pattern noise from the saturation color image, calculate a second fixed pattern noise from the non-saturation color image, generate a final fixed pattern noise by combining the first fixed pattern noise with the second fixed pattern noise, and output a revised image by subtracting the final fixed pattern noise from the raw image.