Focal point detection device and focal point detection method
A focal point detection device comprising: an image sensor having a focal point detection pixel, a pixel data generation section to generate pixel data of a basis part and a reference part corresponding to a predetermined AF area using an output of the focal point detection pixel, a plurality of filter sections having different frequency characteristics to perform filter processing on the pixel data, a detection section to detect contrast of subject image data configured with a plurality of pieces of pixel data of the basis part and the reference part, and a calculation section to select a filter section from among the plurality of filter sections depending on the contrast output by the detection section, to perform correlation calculation using pixel data on which the filter processing is performed by the selected filter section to calculate a defocus amount.
1. A focal point detection device, comprising:
an image sensor having a focal point detection pixel where an incident direction of an incident light flux is restricted and an imaging pixel where the incident direction of the incident light flux is not restricted;
a pixel data generation section to generate focal point detection pixel data of a basis part and a reference part corresponding to a predetermined AF area using an output of the focal point detection pixel;
a plurality of filter sections having different frequency characteristics to perform filter processing on the pixel data;
a detection section to detect contrast of subject image data configured with a plurality of pieces of pixel data of the basis part and the reference part;
a calculation section to select a filter section from among the plurality of filter sections depending on the contrast output by the detection section, to perform correlation calculation using pixel data obtained by performing filter processing on the pixel data of the basis part and the reference part with the selected filter section, and to calculate a defocus amount;
a moving object discrimination section to discriminate whether a subject is a moving object or not, using imaging data output from the imaging pixel; and
a reliability determination section to determine reliability of an output from the calculation section, wherein
the plurality of filter sections includes a first filter processing section having a predetermined frequency characteristic and a second filter processing section having a frequency characteristic which exhibits a larger gain for a low frequency component than the frequency characteristic of the first filter processing section,
the calculation section selects the second filter processing section to perform calculation when the contrast output by the detection section is lower than a predetermined value, and
the calculation section selects the second filter processing section to perform calculation when the moving object determination section determines that the subject is a moving object while the correlation calculation by the calculation section and the reliability determination by the reliability determination section are carried out repeatedly on the basis of pixel data obtained by repeatedly performing imaging operation by the image sensor, and also when the reliability determination section continues to determine that reliability is low for a predetermined time or longer while the calculation section selects the first filter processing section to perform calculation,
wherein the detection section detects an addition value of a difference between a maximum value and a minimum value out of a plurality of pieces of pixel data of the basis part and a difference between a maximum value and a minimum value out of a plurality of pieces of pixel data of the reference part.
2. The focal point detection device according to claim 1 , wherein,
out of outputs from a plurality of the focal point detection pixels corresponding to the basis part and the reference part corresponding to the AF area, the pixel data generation section calculates an average value of outputs from the focal point detection pixels for each arrangement of the focal point detection pixels arranged in a direction perpendicular to an arrangement direction of the basis part and the reference part for each arrangement, to obtain the pixel data.
3. The focal point detection device according to claim 1 , wherein
the reliability determination section obtains a minimum value and correlation values neighboring the minimum value from a plurality of correlation values output by the calculation section, and determines reliability by comparing a larger difference out of differences between the minimum value and the neighboring correlation values with a predetermined threshold value.
4. A focal point detection device, comprising:
an image sensor having a focal point detection pixel where an incident direction of an incident light flux is restricted;
a pixel data generation section to generate focal point detection pixel data of a basis part and a reference part corresponding to a predetermined AF area using an output of the focal point detection pixel;
a plurality of filter sections having different frequency characteristics to perform filter processing on the pixel data;
a detection section to detect contrast of subject image data configured with a plurality of pieces of pixel data of the basis part and the reference part;
a calculation section to select a filter section from among the plurality of filter sections depending on the contrast output by the detection section, to perform correlation calculation using pixel data obtained by performing filter processing on the pixel data of the basis part and the reference part with the selected filter section, and to calculate a defocus amount;
a pan•tilt detection section to detect pan or tilt operation of the focal point detection device; and
a reliability determination section to determine reliability of an output from the calculation section, wherein
the calculation section selects the second filter processing section to perform calculation when the moving object determination section determines that the pan•tilt detection section determines that pan•tilt operation is performed while the correlation calculation by the calculation section and the reliability determination by the reliability determination section are carried out repeatedly on the basis of pixel data obtained by repeatedly performing imaging operation by the image sensor, and also when the reliability determination section continues to determine that reliability is low for a predetermined time or longer while the calculation section selects the first filter processing section to perform calculation,
wherein the detection section detects an addition value of a difference between a maximum value and a minimum value out of a plurality of pieces of pixel data of the basis part and a difference between a maximum value and a minimum value out of a plurality of pieces of pixel data of the reference part.
5. The focal point detection device according to claim 4 , wherein,
out of outputs from a plurality of the focal point detection pixels corresponding to the basis part and the reference part corresponding to the AF area, the pixel data generation section calculates an average value of outputs from the focal point detection pixels for each arrangement of the focal point detection pixels arranged in a direction perpendicular to an arrangement direction of the basis part and the reference part for each arrangement, to obtain the pixel data.
6. The focal point detection device according to claim 4 , wherein
the reliability determination section obtains a minimum value and correlation values neighboring the minimum value from a plurality of correlation values output by the calculation section, and determines reliability by comparing a larger difference out of differences between the minimum value and the neighboring correlation values with a predetermined threshold value.
7. A focal point detection method of a focal point detection device including an image sensor having a focal point detection pixel where an incident direction of an incident light flux is restricted and an imaging pixel where the incident direction of the incident light flux is not restricted, and a plurality of filter sections having different frequency characteristics to perform filter processing on the pixel data, wherein the plurality of filter sections includes a first filter processing section having a predetermined frequency characteristic and a second filter processing section having a frequency characteristic which exhibits a larger gain for a low frequency component than the frequency characteristic of the first filter processing section, the method comprising:
generating focal point detection pixel data of a basis part and a reference part corresponding to a predetermined AF area using an output of the focal point detection pixel;
detecting contrast of subject image data configured with a plurality of pieces of pixel data of the basis part and the reference part;
selecting a filter section from among the plurality of filter sections depending on the detected contrast;
performing correlation calculation using pixel data obtained by performing filter processing on the pixel data of the basis part and the reference part with the selected filter section, to calculate a defocus amount; and
selecting the second filter processing section to perform calculation when a subject is determined to be a moving object on the basis of imaging data output from the imaging pixel while the correlation calculation and reliability determination of the correlation calculation are carried out repeatedly on the basis of pixel data obtained by repeatedly performing imaging by the image sensor, and also when it is continued to be determined that reliability is low in the result of operation for a predetermined time or longer while the first filter processing section is selected and the calculation is performed,
wherein an addition value of a difference between a maximum value and a minimum value out of a plurality of pieces of pixel data of the basis part and a difference between a maximum value and a minimum value out of a plurality of pieces of pixel data of the reference part is detected.
8. The focal point detection method according to claim 7 , wherein,
in generation of the pixel data, out of outputs from a plurality of focal point detection pixels in the basis part and the reference part corresponding to the AF area, an average value of outputs from the focal point detection pixels of each arrangement is calculated for each arrangement of the focal point detection pixels arranged in a direction perpendicular to an arrangement direction of the basis part and the reference part, to give the pixel data.
9. The focal point detection method according to claim 7 , wherein,
in the reliability determination, a minimum value and correlation values neighboring the minimum value are obtained from a plurality of correlation values calculated from the correlation calculation, and reliability is determined by comparing a larger difference out of differences between the minimum value and the neighboring correlation values with a predetermined threshold value.
10. A focal point detection method of a focal point detection device including an image sensor having a focal point detection pixel where an incident direction of an incident light flux is restricted, a plurality of filter sections having different frequency characteristics to perform filter processing on the pixel data, wherein the plurality of filter sections includes a first filter processing section having a predetermined frequency characteristic and a second filter processing section having a frequency characteristic which exhibits a larger gain for a low frequency component than the frequency characteristic of the first filter processing section, and a pan•tilt detection section to detect pan or tilt operation of the focal point detection device, the method comprising:
generating focal point detection pixel data of a basis part and a reference part corresponding to a predetermined AF area using an output of the focal point detection pixel;
detecting contrast of subject image data configured with a plurality of pieces of pixel data of the basis part and the reference part;
selecting a filter section from among the plurality of filter sections depending on the detected contrast;
performing correlation calculation using pixel data obtained by performing filter processing on the pixel data of the basis part and the reference part with the
selected filter section, to calculate a defocus amount; and
selecting the second filter processing section to perform the correlation calculation when the pan•tilt detection section determines that a pan•tilt operation is performed while the correlation calculation and reliability determination of the correlation calculation are carried out repeatedly on the basis of pixel data obtained by repeatedly performing imaging by the image sensor, and also when it is continued to be determined that the reliability is low for a predetermined time or longer while the first filter processing section is selected and the correlation calculation is performed,
wherein an addition value of a difference between a maximum value and a minimum value out of a plurality of pieces of pixel data of the basis part and a difference between a maximum value and a minimum value out of a plurality of pieces of pixel data of the reference part is detected.
11. The focal point detection method according to claim 10 , wherein,
in generation of the pixel data, out of outputs from a plurality of focal point detection pixels in the basis part and the reference part corresponding to the AF area, an average value of outputs from the focal point detection pixels of each arrangement is calculated for each arrangement of the focal point detection pixels arranged in a direction perpendicular to an arrangement direction of the basis part and the reference part, to give the pixel data.
12. The focal point detection method according to claim 10 , wherein
in the reliability determination, a minimum value and correlation values neighboring the minimum value are obtained from a plurality of correlation values calculated from the correlation calculation, and reliability is determined by comparing a larger difference out of differences between the minimum value and the neighboring correlation values with a predetermined threshold value.