Video Image Anti-Shake With Z-Axis Translation Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image stabilization technologies, such as OIS and EIS, fail to effectively compensate for translational shake on the Z-axis and do not consider object distance, leading to poor stabilization and high power consumption.
Innovation Solution
A video image anti-shake method that detects shake on the X, Y, and Z-axes, including object distance and image distance, and performs anti-shake processing using sensors like gyroscopes, accelerometers, and depth sensors to compensate for six-axis shake, adjusting image stabilization based on detected shake types and distances.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If image-content-based anti-shake processing is used to identify motion situation and perform cropping, stretching, and deformation, then anti-shake effect is improved, but calculation amount increases, processing speed decreases, and power consumption increases
Solution Approach 1:
The patent segments the anti-shake processing into two distinct pathways: motion sensor-based processing for dynamic scenes and image-content-based processing for static scenes. This segmentation allows the system to use the low-power motion sensor pathway whenever possible, only activating the high-power image-content-based pathway when necessary, thereby resolving the contradiction between anti-shake effectiveness and power consumption.
Solution Approach 2:
The patent dynamically switches between different anti-shake processing modes based on scene characteristics. The system continuously monitors motion sensor data and transitions between motion sensor-based processing and image-content-based processing, optimizing the balance between stabilization quality and energy consumption in real-time.
2Productivity
If motion sensor data is used to perform anti-shake processing based on rotation and translation, then processing speed is improved and power consumption is reduced, but compensation for Z-axis translational shake is not implemented
Solution Approach 1:
The patent extends the motion sensor-based processing to handle all six degrees of freedom including Z-axis translational shake. The same motion sensor data is processed through multiple calculation pathways to compensate for rotation, translation, and Z-axis movement, making the motion sensor-based system universally applicable to all shake types rather than limited to only X-Y plane movements.
Solution Approach 2:
The patent adds the Z-axis dimension to the traditional two-dimensional (X-Y plane) anti-shake processing. By incorporating depth information and calculating image displacement in the Z-direction based on motion sensor data, the system transitions from planar stabilization to three-dimensional stabilization, comprehensively compensating for all spatial shake components.
3Measurement precision
If object distance detection is added to improve anti-shake accuracy, then stabilization precision is improved, but device complexity increases
Solution Approach 1:
The patent uses object distance as an intermediary parameter that bridges motion sensor data and image stabilization. Rather than directly measuring image displacement, the system uses motion sensor data combined with object distance information to calculate expected image displacement, providing a more accurate reference for stabilization while avoiding the need for complex direct measurement systems.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Achieves accurate stabilization of video images by compensating for translational shake on all axes, reducing data processing volume and power consumption, and improving image quality during focus adjustments.
Implementation Method 1
detecting rotation shake on the X-axis, the Y-axis, and the Z-axis
Implementation Method 2
detecting translational shake on the X-axis and the Y-axis
Implementation Method 3
detecting an object distance in a photographing scenario, where the object distance is a distance to a focused object or person
Data Source
Figure 1
Figure 2(a)~2(b)
Figure 3
AI summary
This application discloses a video image anti-shake method and a terminal, and relates to the field of image processing, to implement compensation for translational shake on a Z direction. A video image anti-shake method includes: turning on, by a terminal, a camera lens, and photographing a video image by using the camera lens; detecting, by the terminal, shake on an X-axis, a Y-axis, and a Z-axis during photographing, where the Z-axis is an optical axis of the camera lens, the X-axis is an axis perpendicular to the Z-axis on a horizontal plane, and the Y-axis is an axis perpendicular to the Z-axis on a vertical plane; and performing, by the terminal, anti-shake processing on the video image based on the shake on the X-axis, the Y-axis, and the Z-axis. Embodiments of this application are applied to video image anti-shake.