Self-Leveling Pipe Inspection Camera for Low-Latency Video

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current self-leveling pipe inspection systems face challenges in maintaining real-time video feed quality due to high memory resource consumption, often resulting in latency, cropping, or rescaling of images, which compromises image quality and field of view.

Innovation Solution

A self-leveling pipe inspection system with a camera head equipped with an orientation sensing module and image processing module that generates a subset of pixel values based on orientation signals, allowing for efficient image processing and transmission while maintaining image quality and field of view.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional digital rotation transformation is applied to rotate images in real-time, then image orientation is corrected, but memory resource consumption increases significantly causing unacceptable latency

Engineering Contradiction:
Improveimage orientation accuracyVSAvoidvideo latency
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent extracts only the essential orientation correction information from the full image transformation process. Instead of applying complete digital rotation transformation to the entire image, the system extracts orientation data from accelerometer sensors and applies minimal processing to generate corrected video frames, significantly reducing memory resource consumption and latency while maintaining orientation accuracy.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent applies partial action by performing only the necessary orientation correction operations rather than complete image rotation. The system uses sensor data to determine the degree of orientation adjustment needed and applies only that specific correction, avoiding unnecessary computational overhead and memory usage associated with full digital rotation transformation.

Inventive Principle:
Principle #16Partial or excessive action

2Loss of time

If memory resource consumption is reduced through cropping or rescaling, then video latency decreases, but image quality and field of view are compromised

Engineering Contradiction:
Improvevideo latencyVSAvoidimage quality
Core Design Contradiction:
Loss of timeVSManufacturing precision

Solution Approach 1:

The patent applies preliminary action by pre-processing the image data to extract and prioritize essential visual information before transmission. The system uses orientation sensors to pre-determine which portions of the image require full resolution and which can be processed more efficiently, allowing real-time orientation correction without compromising overall image quality or increasing latency.

Inventive Principle:
Principle #10Preliminary action

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

The system provides real-time, orientation-adjusted video feeds with improved memory efficiency, reducing latency and preserving image quality and field of view, enhancing overall inspection system performance.

Implementation Method 1

generating an orientation signal corresponding to an orientation of the image sensor, which may be a gravitation orientation

Methodology Applied
Scientific EffectGravitation: Gravitation

Data Source

PatentUS11674906B1Self-leveling inspection systems and methods
Publication Date: 2023.06.13 SEESCAN INC
  • US11674906B1 patent drawing
  • US11674906B1 patent drawing
  • US11674906B1 patent drawing

AI summary

Self-leveling inspection system methods and devices for use in inspecting buried pipes or other cavities are disclosed. The system includes a camera head with an image sensor, an orientation sensor, and a signal processing module including a processing programmed to receive an image from the image sensor, and an orientation signal from the orientation sensor, generate a second image based at least in part on information provided from the orientation sensor, and store the second image in a non-transitory memory.