US 10529029

Summary of US Patent 10529029

Patent Number: US10529029B2
Title: Platform, systems, and methods for identifying property characteristics and property feature maintenance through aerial imagery analysis
Assignee: Aon Benfield Inc.
Inventor(s): Takeshi Okazaki
Filing Date: September 25, 2017
Issue Date: January 7, 2020

Abstract:
The patent describes methods and systems for automatically assessing the repair condition of a property characteristic. This involves obtaining aerial imagery of a property's geographic region, identifying features within that imagery corresponding to a specific property characteristic, and then analyzing those features to classify the characteristic itself (e.g., roof shape). Subsequently, a region of the aerial image around that characteristic is analyzed to determine its condition (e.g., good or bad). Finally, based on both the characteristic classification and the condition classification, the system determines either a risk estimate of damage to the property from disasters or a cost estimate for repairing or replacing the characteristic.

---

Plain-Language Overview of Independent Claims:

Here's an overview of the core independent claims as derived from the "SUMMARY OF ILLUSTRATIVE EMBODIMENTS" section of the patent:

1. Method for Categorizing Repair Condition and Estimating Damage Risk: This claim describes a computer-implemented method. It involves:

   • Getting an aerial image of a property.
   • Finding specific features in that image that relate to a property characteristic (like a roof).
   • Analyzing these features to identify the type of property characteristic (e.g., classifying the roof as a gable or hipped roof).
   • Analyzing the part of the image showing the characteristic to determine its condition (e.g., whether the roof is in good or poor condition).
   • Using both the type of characteristic and its condition to calculate how much risk the property faces from one or more disasters.

2. System for Categorizing Repair Condition and Estimating Replacement Cost: This claim outlines a system that includes computer hardware (processing circuitry) and software instructions. When the software runs, it causes the system to:

   • Obtain an aerial image of a property.
   • Identify features in the image related to a specific property characteristic.
   • Analyze these features to classify the type of property characteristic.
   • Analyze the region of the image containing the characteristic to determine its condition.
   • Use the characteristic's type and condition to calculate the cost to replace that property characteristic.

3. Non-Transitory Computer-Readable Medium for Risk Estimation: This claim covers a computer storage medium containing instructions. When a computer's processing hardware executes these instructions, it enables the computer to:

   • Receive information identifying a property and at least one of its characteristics.
   • Obtain an aerial image of the property's location.
   • Identify features in the aerial image corresponding to each specified property characteristic.
   • For each property characteristic, analyze its features to classify the characteristic and analyze the image region to determine its condition.
   • Determine at least one risk estimate for damage due to disaster, based on both the classification and condition of each property characteristic.

CAFC 2026 Dockets:
A review of CAFC 2026 dockets did not identify any specific litigation directly involving US patent 10529029. While general patent eligibility cases concerning machine learning were noted, no direct correlation to this patent number was found within the search results for the specified timeframe.