KPNDepth: Depth Estimation of Lane Images under Complex Rainy Environment

TL;DR

This paper presents KPNDepth, a neural network for accurate lane depth estimation in rainy conditions, utilizing a dual-layer kernel prediction approach and a synthetic rainy dataset to improve robustness.

Contribution

Introduction of a dual-layer convolutional kernel prediction network and a synthetic rainy dataset for enhanced lane depth estimation in rainy environments.

Findings

Effective in complex rainy conditions

Outperforms existing methods in rainy scenarios

Synthetic dataset improves training and evaluation

Abstract

Recent advancements in deep neural networks have improved depth estimation in clear, daytime driving scenarios. However, existing methods struggle with rainy conditions due to rain streaks and fog, which distort depth estimation. This paper introduces a novel dual-layer convolutional kernel prediction network for lane depth estimation in rainy environments. It predicts two sets of kernels to mitigate depth loss and rain streak artifacts. To address the scarcity of real rainy lane data, an image synthesis algorithm, RCFLane, is presented, creating a synthetic dataset called RainKITTI. Experiments show the framework's effectiveness in complex rainy conditions.