RASTER-TO-VECTOR CONVERSION OPERATIONS ADAPTED TO MODELING OF RF PROPAGATION

Patent №

US 7,047,014

Granted

2006-05-16

Filed 2004

Owner

AIRESPACE, INC.

+1 more

Lab

AI components

2

nlp · vision

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

10982153

Methods, apparatuses and systems directed to optimizing vector models for use in modeling RF propagation in desired physical environments. In one embodiment, the present invention can operate on pre-existing vector models. In other implementations, the present invention facilitates the conversion of raster images of buildings and other physical locations to vector formats for use in connection with the computational modeling of radio-frequency (RF) propagation. According to certain embodiments, the present invention is implemented within the context of a location diagram editing application that supports a line recognition filter, a snap filter and a merge filter which a user may individually select and configure. As discussed more fully below, the line recognition filter operates on the vector objects to adjust near-vertical lines (as defined by a configurable threshold angle) to vertical, and near-horizontal lines to horizontal. In some configuration modes, the line recognition filter may also delete lines that are oriented at an angle greater than a threshold from the x-axis and y-axis to, for example, clean up vector objects corresponding to text and other elements that are usually analytically irrelevant to modeling RF propagation. The snap filter, in one implementation, operates to extend closely spaced objects that are oriented within a threshold distance from each other. The merge filter operates, in one implementation, to merge closely spaced parallel lines to, for example, adjust the vector file to correctly model a wall with a single vector object. In one implementation, the output resulting from the invention is a vector file that has each wall accorded a given wall type, and is optimized for use in connection with an RF prediction engine.

Natural languageVisionH04W 16/18H04W 4/18H04W 64/00

AI classification

Vision0.97
Natural language0.74
AI hardware0.17
Machine learning0.11
Knowledge representation0.00
Speech0.00
Planning0.00
Evolutionary computation0.00

Ownership

AIRESPACE, INC.

assignment · 159670606

CISCO SYSTEMS, INC.

merger · 211380063

Assignors

FRIDAY, ROBERT J., DIETRICH, PAUL F., DAVI, GREGG SCOTT

On an employer assignment, the assignors are typically the inventors.

© 2026 NYSGPT2525 LLC