Dynamic Interference Suppression in Surveillance Object Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods fail to accurately distinguish and suppress interfering objects, such as moving vehicle doors, from target objects in image sequences, leading to difficulties in detecting and counting target objects in monitoring regions like vehicle entrances.
Innovation Solution
A method that adjusts pixel values in image sequences by accounting for interference components from known interfering objects, using a movement model to simulate and eliminate interference, allowing for accurate separation and detection of target objects in a three-dimensional coordinate system.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image processing methods are used to suppress stationary interfering objects, then target object detection is improved, but moving interfering objects (like vehicle doors) cannot be suppressed because they do not fit the stationary object assumption
Solution Approach 1:
The patent extends the suppression method from stationary to moving interfering objects by dynamically tracking and modeling the motion of interfering objects. The system maintains a database of interfering objects with their motion characteristics and updates their positions and movement patterns in real-time, allowing the suppression algorithm to adapt to dynamic scenes where interfering objects move through the monitoring region.
Solution Approach 2:
The patent changes the parameters used for object classification from static (stationary vs. moving) to dynamic motion characteristics. By analyzing motion vectors, speed, direction, and trajectory of objects, the system can distinguish between target objects and moving interfering objects based on their motion patterns rather than just their presence or absence in the image.
2Device complexity
If moving interfering objects are not suppressed, then the detection method remains simple, but detection accuracy deteriorates due to false positives from interfering objects
Solution Approach 1:
The patent performs preliminary actions by pre-identifying and database-ing interfering objects before the actual detection process. The system maintains a database of interfering objects with their typical motion patterns and positions, allowing the detection algorithm to preemptively filter out these objects based on their known characteristics before analyzing potential target objects.
Solution Approach 2:
The patent implements feedback mechanisms where the detection system continuously monitors and learns from the behavior of interfering objects. By analyzing the motion patterns of known interfering objects and comparing them with detected objects, the system refines its classification criteria and improves its ability to distinguish between targets and interferers over time.
Data Source
AI summary
The present application presents methods and apparatuses for detecting target objects in an image sequence of a monitoring region. In some examples, such methods may include adjusting pixel values of images of the image sequence for interference components associated with at least one interfering object, generating the interference components associated with the at least one interfering object that is situated in the monitoring region, searching the image sequence for the target objects based on the adjusted pixel values, detecting a start of a predetermined sequence of motions associated with the interfering object, and computing an instantaneous position of the at least one interfering object during the predetermined sequence of motions, wherein adjusting the pixel values of the images is based upon the instantaneous position.