Image processing apparatus adapted to recognize object in acquired image
First Claim
1. An image processing apparatus comprising:
- a processor configured to at least perform;
an image data acquisition for acquiring, from an image acquisition apparatus provided with an image acquisition region comprising both a light-shielded region where a plurality of light-shielded solid-state image acquisition elements are arrayed as light-shielded pixels and a non-light-shielded region where a plurality of exposed solid-state image acquisition elements are arrayed two-dimensionally as effective pixels for acquiring an image of an object, data of the image based on output signals from the effective pixels which are provided by the solid-state image acquisition elements in the non-light-shielded region of the image acquisition apparatus;
an output acquisition for acquiring output signals from the light-shielded pixels which are provided by the solid-state image acquisition elements in the light-shielded region;
a noise estimation for estimating a degree of variations in levels of noise contained in the output signals from the effective pixels, based on the output signals acquired by the output acquisition;
a threshold setting for setting the edge threshold based on both the degree of variations in noise which is estimated by the noise estimation and a maximum contrast which is set as a difference between a maximum and a minimum among levels of the output signals from the effective pixels based on the data acquired by the image acquisition;
an edge extraction for extracting an edge in the image among the data of the image acquired by the image data acquisition, the edge being composed of pixels whose values has a difference equal to or large than the edge threshold between adjacent pixels in the effective pixels, the edge threshold being set by the threshold setting; and
an object recognition for recognizing the object based on the edge extracted by the edge extraction,wherein the threshold setting is configured to set the edge threshold at a value falling into a range i) which is lower than the maximum contrast and ii) which allows the levels of the noise exceeding the edge threshold to occur at a probability lower than a preset value.
1 Assignment
0 Petitions
Accused Products
Abstract
An image processing apparatus communicates with an image acquisition apparatus provided with an image acquisition region comprising light-shielded pixels and effective pixels. Data of an image are acquired based on output signals from the effective pixels. An edge of an object is extracted in the acquired image data using a preset edge threshold, and the object is recognized based on the extracted edge. Output signals are acquired from the light-shielded pixels and a degree of variations in noise contained in the output signals from the effective pixels is estimated based on the output signals acquired. The edge threshold is set based on the degree of variations in noise which is estimated, such that the noise having a level which exceeds the edge threshold occurs at a probability lower than a preset value.
-
Citations
9 Claims
-
1. An image processing apparatus comprising:
a processor configured to at least perform; an image data acquisition for acquiring, from an image acquisition apparatus provided with an image acquisition region comprising both a light-shielded region where a plurality of light-shielded solid-state image acquisition elements are arrayed as light-shielded pixels and a non-light-shielded region where a plurality of exposed solid-state image acquisition elements are arrayed two-dimensionally as effective pixels for acquiring an image of an object, data of the image based on output signals from the effective pixels which are provided by the solid-state image acquisition elements in the non-light-shielded region of the image acquisition apparatus; an output acquisition for acquiring output signals from the light-shielded pixels which are provided by the solid-state image acquisition elements in the light-shielded region; a noise estimation for estimating a degree of variations in levels of noise contained in the output signals from the effective pixels, based on the output signals acquired by the output acquisition; a threshold setting for setting the edge threshold based on both the degree of variations in noise which is estimated by the noise estimation and a maximum contrast which is set as a difference between a maximum and a minimum among levels of the output signals from the effective pixels based on the data acquired by the image acquisition; an edge extraction for extracting an edge in the image among the data of the image acquired by the image data acquisition, the edge being composed of pixels whose values has a difference equal to or large than the edge threshold between adjacent pixels in the effective pixels, the edge threshold being set by the threshold setting; and an object recognition for recognizing the object based on the edge extracted by the edge extraction, wherein the threshold setting is configured to set the edge threshold at a value falling into a range i) which is lower than the maximum contrast and ii) which allows the levels of the noise exceeding the edge threshold to occur at a probability lower than a preset value. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9)
Specification