Image Processor Fusion for Dense, Accurate LiDAR Depth Maps

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing light detection and ranging (LiDAR) systems produce sparse depth maps with high accuracy but low density, limiting their application, while image sensors offer wider use but low accuracy, necessitating a solution for generating dense depth maps with high accuracy and high density.

Innovation Solution

An image processor comprising a first pre-processor to generate a first point cloud from a sparse depth map, a first post-processor to model the surface, a second pre-processor to create a difference frame, a second post-processor to generate a second point cloud, and a generator to produce a dense depth map using the modeled surface and second point cloud.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If LiDAR is used to acquire depth information, then measurement precision is improved, but quantity of substance (density of depth map) deteriorates

Engineering Contradiction:

Improvedepth map accuracyVSAvoiddepth map density

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent merges LiDAR and image sensor data by generating a difference frame from optical flow between consecutive frames, then combining this with the sparse depth map from LiDAR to produce a dense depth map that has both high accuracy and high density

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The difference frame acts as an intermediary that bridges the sparse LiDAR depth map and the dense image sensor data, enabling the transfer of depth information to create a dense depth map with high accuracy

Inventive Principle:

Principle #24Intermediary (Mediator)

2Adaptability or versatility

If image sensor is used to capture images, then adaptability or versatility is improved, but measurement precision deteriorates

Engineering Contradiction:

Improveweather and situation rangeVSAvoiddepth map accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent combines image sensor data with LiDAR depth information by processing optical flow to generate difference frames, which are then used to create a dense depth map that maintains the versatility of image sensors while achieving high measurement precision through LiDAR integration

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS20250245843A1Image processor and image processing system including the same

Publication Date: 2025.07.31 SK HYNIX INC

AI summary

Disclosed is an image processor and an image processing system including the same, and the image processor may include a first pre-processor configured to generate a first point cloud based on a sparse depth map, a first post-processor configured to model a surface of the sparse depth map based on the first point cloud, a second pre-processor configured to generate a difference frame corresponding to a difference between a previous frame and a current frame, a second post-processor configured to generate a second point cloud of a dimension corresponding to the sparse depth map, based on the difference frame, and a generator configured to generate a dense depth map using the second point cloud and the modeled surface.