Image Capture Apparatus Focus Correction Using Peripheral Defocus Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Image capture apparatuses face challenges in maintaining focus detection accuracy when subjects at different distances are included in the autofocus area, leading to reduced pixel count and decreased accuracy and reliability.
Innovation Solution
An image capture apparatus and method that detect a subject area and a characteristic area within it, set an autofocus area, detect defocus amounts for the autofocus area and its peripherals, correct the defocus amount based on peripheral defocus amounts, and adjust the focus of the imaging optical system accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the AF area is reduced to exclude unintended subjects at different distances, then the possibility of unintended subjects being included is reduced, but the pixel count of the AF area is reduced leading to decreased accuracy and reliability of focus detection
Solution Approach 1:
The patent divides the focus detection process into two segments: primary focus detection within the characteristic area (AF area) and secondary reference detection in peripheral areas. By segmenting the detection zones and using peripheral areas as reference points, the system can correct defocus amounts without increasing the AF area size, thus maintaining reliability while preserving measurement precision.
Solution Approach 2:
The patent introduces peripheral areas as intermediary reference zones that mediate between the limited AF area and the full image. These peripheral areas serve as intermediaries to provide additional depth information that compensates for the small pixel count in the AF area, thereby improving accuracy without compromising reliability.
2Measurement precision
If the AF area is set to a characteristic area within a subject area for precise focus adjustment, then focus accuracy for the characteristic area is improved, but subjects at different distances may be included reducing detection accuracy
Solution Approach 1:
The patent segments the image into the AF area (characteristic area) and peripheral areas, allowing separate analysis of each zone. This segmentation enables the system to maintain a small AF area for precise characteristic area focus while using peripheral areas to verify and correct depth information, thus preserving both measurement precision and reliability.
Solution Approach 2:
The patent implements a feedback mechanism where defocus amounts detected in peripheral areas are used to correct the defocus amount detected in the AF area. This feedback loop ensures that even when subjects at different distances are present, the final focus detection remains accurate and reliable by continuously refining the measurement based on multiple reference points.
3Measurement precision
If the AF area is enlarged to increase pixel count for better focus detection, then the accuracy of focus detection is improved, but unintended subjects at different distances are more likely to be included
Solution Approach 1:
The patent extracts reference information from peripheral areas outside the AF area to compensate for the limited pixel count within the AF area. By taking out additional depth cues from the periphery and applying them as corrections, the system achieves high measurement precision without enlarging the AF area, thus avoiding the harmful effect of including unintended subjects.
Data Source
AI summary
An image capture apparatus that detects, from an image, a subject area and a characteristic area within the subject area is disclosed. The image capture apparatus sets an AF area to the characteristic area and detects a defocus amount for the AF area and defocus amounts for peripheral areas of the AF area. The image capture apparatus corrects the defocus amount detected for the AF area based on the defocus amounts detected for the peripheral areas and adjusts focus of an imaging optical system based on the corrected defocus amount.