Gimbal Key-Point Tracking for Precise Partial Close-Up Following
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing gimbal control methods are inadequate for precisely following a partial close-up of a target object, particularly at close distances, leading to issues like image shaking, unsmooth switching, and defocusing.
Innovation Solution
A gimbal control method that detects human body key points in a shot image using skeletal point detection and Convolutional Neural Networks, allowing users to select and adjust the gimbal's attitude and shooting parameters based on these key points to achieve smooth and intelligent focus adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If full body or half body following is used, then the method is suitable for long-distance or medium-distance scenes, but it cannot precisely follow a partial close-up
Solution Approach 1:
The patent segments the target object into multiple body parts by detecting key points (head, chest, hands, feet). Instead of treating the entire body as a single following target, the system divides it into selectable segments, allowing precise following of specific body parts while maintaining adaptability across different shooting scenarios.
Solution Approach 2:
The patent adds a new dimension of control by introducing body part selection beyond the traditional full-body following mode. Users can now select specific body parts (head, chest, hands, feet) as following targets, transforming a single-dimensional following approach into a multi-dimensional one that includes both distance and body part dimensions.
2Reliability
If traditional following method is used, then the operation is simple, but the photographing effect is not optimized and image blurring occurs
Solution Approach 1:
The system performs automatic body key point detection and provides intelligent following control without requiring manual operation. The gimbal automatically detects body parts, calculates optimal following parameters, and adjusts shooting parameters to prevent image blurring, making the complex operations transparent to the user while maintaining simple interaction.
Solution Approach 2:
The system continuously monitors body key point positions and uses this feedback to dynamically adjust the following control. By real-time detection of body part movements and feedback-based adjustment of gimbal parameters, the system optimizes photographing quality and prevents image blurring while maintaining operational simplicity.
3Ease of operation
If manual adjustment is used, then the control is straightforward, but the user experience is poor and learning complexity is high
Solution Approach 1:
The system performs preliminary action by automatically detecting body key points and pre-calculating optimal following parameters before the user needs to adjust anything. This preliminary detection and calculation eliminates the need for manual adjustment operations, reducing both operational complexity and adjustment time while improving user experience.
Data Source
AI summary
A control method includes obtaining one or more target body key points from a shot image collected by a shooting apparatus carried by a gimbal and controlling the gimbal to adjust an attitude and/or controlling the shooting apparatus to adjust a shooting parameter according to the one or more target body key points. The one or more target body key points are used to indicate a body part of the target object. The gimbal is configured to carry the shooting apparatus and drive the shooting apparatus to rotate to adjust the attitude of the shooting apparatus.