A method and device for detecting foot points of moving obstacles in an aerial view using frame difference method

By using the frame difference method for obstacle detection in bird's-eye view, and by utilizing the grayscale and ternary values ​​of multiple consecutive panoramic images, combined with the differential histogram and preset threshold, the problem of detecting obstacle foot points when the camera is moving is solved, thus achieving fast and accurate obstacle detection.

CN115620262BActive Publication Date: 2026-06-16WUHAN KOTEI INFORMATICS
2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
WUHAN KOTEI INFORMATICS
Filing Date
2022-10-17
Publication Date
2026-06-16

AI Technical Summary

⚠Technical Problem

Existing frame difference methods have difficulty effectively detecting moving obstacles when the camera is moving, especially during automatic parking, where they cannot accurately calculate the foot points of moving obstacles.

⚗Method used

A method for detecting the foot points of moving obstacles in a bird's-eye view using the frame difference method is proposed. This method acquires multiple consecutive panoramic images, performs grayscale and ternary conversion, draws a difference histogram from multiple perspectives, uses the difference histogram and a preset proportion threshold to determine the foot points of the obstacles, and calculates the final foot point by angle and distance.

🎯Benefits of technology

It achieves accurate detection of obstacle foot points even when the camera is moving, with low computational load and high speed, making it suitable for automatic parking scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
Patent Text Reader

Abstract

The application relates to a bird's-eye-view moving obstacle foot point detection method and device based on a frame difference method, which comprises the following steps: acquiring multiple frames of continuous ring-view images containing at least one moving obstacle, and sequentially performing gray-scale processing and three-value processing on the images; based on the multiple frames of three-value processed continuous ring-view images, a multi-view difference histogram is drawn; according to the difference histogram and a preset proportion threshold of a target obstacle, multiple foot points of each moving obstacle are determined; and through angle and distance calculation, the multiple foot points are used to obtain a final foot point of each moving obstacle. The application utilizes multiple frames of continuous ring-view images, reduces the number of cameras, solves the problem that a traditional frame difference method must be fixed in the background, and realizes detection of the moving obstacle, so that the method is simple, the calculation amount is small, and the speed is fast.
Need to check novelty before this filing date? Find Prior Art