Flying Spot Detection in Depth Images via Normal Vector Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for detecting flying spots in depth images using TOF technology suffer from high error rates and affect image quality due to difficulties in adjusting pixel width, leading to removal of effective pixels and blurring of structural features.
Innovation Solution
A method that calculates normal vectors for each pixel in a depth image, uses an edge detection operator to identify edge points, and determines seed points as flying spots based on a growing condition, thereby accurately detecting and removing flying spots.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If convolution operation with operators such as sobel, canny, or Laplace is executed on the image to obtain a contour of the edge, then the edge contour can be detected, but the pixel width of the contour is difficult to adjust effectively, causing some effective pixels to be removed by mistake
Solution Approach 1:
The patent changes the detection parameters from fixed operator thresholds to dynamically adjustable parameters including pixel width W, growth threshold T1, and termination threshold T2. These parameters can be adjusted according to different image characteristics and requirements, allowing flexible control over detection sensitivity and pixel removal criteria.
Solution Approach 2:
The patent introduces a dynamic growing process where the detection region expands iteratively from seed points based on grayscale differences. The growth continues until the grayscale difference exceeds threshold T2, creating an adaptive detection boundary that adjusts to local image characteristics rather than using fixed contour lines.
2Object-affected harmful factors
If pixels falling within the contour are regarded as flying spots and removed, then the flying spot effect can be reduced, but the error rate is excessively high and structural features of the image are affected, resulting in blurring of details
Solution Approach 1:
The patent segments the image into different regions based on edge detection results, identifying specific flying spot areas rather than removing all pixels within a broad contour. This selective segmentation allows preservation of important structural features while removing only the harmful flying spot artifacts.
Solution Approach 2:
The patent uses a feedback mechanism where the detection process iteratively refines the identification of flying spots by comparing grayscale differences against adaptive thresholds. The system continuously adjusts based on detected patterns, improving accuracy and reducing false removal of effective pixels.
3Difficulty of detecting and measuring
If the contour obtained by calculating with operators such as sobel, canny, etc. is used, then the edge can be identified, but the pixel width is difficult to adjust effectively with parameters of operators
Solution Approach 1:
The patent replaces fixed operator parameters with a flexible parameter set including pixel width W, growth threshold T1, and termination threshold T2. These parameters can be independently adjusted to adapt to different image scales, resolutions, and content types, providing superior versatility compared to traditional operators.
Solution Approach 2:
The patent transitions from 2D contour detection to a 3D-like growing process by introducing the pixel width dimension W and iterative growth steps. This adds a temporal/iterative dimension to the detection process, allowing more nuanced control over detection behavior.
Data Source
AI summary
Provided are a method and a device for detecting a flying spot on an edge of a depth image and an electronic device. The method includes the following. Image data of the depth image is collected. A normal vector of each pixel of the depth image is calculated according to the image data. An edge point of the depth image is calculated according to the normal vector with an edge detection operator. The edge point is taken as a seed point, and whether the seed point satisfies a growing condition is determined. Upon determining that the seed point satisfies the growing condition, the seed point is determined as the flying spot.