Lens Control Apparatus Using Kalman Filter Selection for Moving Object Focus
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for predicting the image plane position of a moving object, such as those using the batch least squares method, face challenges in achieving accurate and efficient calculations, particularly when dealing with a large number of past data points, leading to increased processing loads and potential errors in focus detection.
Innovation Solution
The proposed solution incorporates a digital single lens reflex camera system with a focus detection unit, an estimation unit using sequential identification methods like the Kalman filter, and a prediction unit that selects between different estimation methods based on object motion and convergence criteria to predict future image plane positions, thereby controlling the focus lens effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If batch least squares method is used to predict image plane position, then prediction accuracy can be improved, but processing load increases significantly when dealing with large number of past data points
Solution Approach 1:
The patent divides the estimation process into multiple independent estimation units (first estimation unit, second estimation unit, third estimation unit), each handling different aspects of image plane position prediction. This segmentation allows parallel processing and reduces the computational burden on any single unit, thereby lowering overall processing load while maintaining prediction accuracy.
Solution Approach 2:
The patent dynamically selects which estimation unit to use based on the motion state of the object. When the object is stationary or moving slowly, simpler estimation methods are employed. When the object is moving rapidly, more complex estimation methods are activated. This dynamic adaptation optimizes processing load according to actual needs, avoiding unnecessary computational overhead.
2Measurement precision
If batch least squares method is used with large number of past data points, then prediction accuracy improves, but calculation time increases
Solution Approach 1:
The patent uses multiple estimation units that can operate independently and partially. Each estimation unit processes a subset of the data or a specific aspect of the prediction, allowing the system to achieve adequate prediction accuracy without processing all possible data points through the most computationally intensive methods, thus reducing calculation time.
Solution Approach 2:
The patent performs preliminary estimation using simpler methods in earlier estimation units before proceeding to more complex calculations. This preliminary action filters out obvious cases and prepares data in advance, reducing the computational workload for subsequent more accurate but time-consuming estimation processes.
3Device complexity
If single estimation method is used, then device complexity is reduced, but adaptability to different object motion states decreases
Solution Approach 1:
The patent creates a universal estimation system that can handle multiple types of object motion states through multiple estimation units. Each unit is designed to handle specific motion characteristics, and the control unit selects the appropriate unit based on detected motion state, providing universal adaptability across different scenarios while maintaining relatively simple individual unit designs.
Solution Approach 2:
The patent implements dynamic selection among different estimation units based on the motion state of the object. The control unit monitors object motion and dynamically switches between estimation methods, enabling the system to adapt to varying motion conditions without requiring a completely different system design for each scenario.
Data Source
AI summary
A microcomputer estimates information corresponding to an image plane position of an object based on a sequential identification method using a focus detection result. Moreover, the microcomputer predicts information corresponding to a future image plane position of the object based on the estimated information corresponding to the image plane position of the object. The microcomputer controls driving of a focus lens based on the predicted information corresponding to the future image plane position of the object. In accordance with information corresponding to a motion of the object, the microcomputer selects by which of a first Kalman filter and a second Kalman filter the information corresponding to the future image plane position of the object is predicted by using the estimated information corresponding to the image plane position of the object.