ECSNet: An Accelerated Real-Time Image Segmentation CNN Architecture for Pavement Crack Detection

The ability to perform pixel-wise segmentation on pavement cracks in real-time is paramount in road service condition assessment and maintenance decision-making practices. Recent deep learning detection models are focused on detection accuracy and require a large number of computing sources and long run times. However, highly efficient and accelerated models with acceptable accuracy in real-time pavement crack detection tasks are required but hard to achieve. In this work, we present a customized deep learning model architecture named Efficient Crack Segmentation Neural Network (ECSNet) for accelerated real-time pavement crack detection and segmentation without compromising performance. We introduce some novel parts, including small kernel convolutional layers and parallel max pooling and convolutional operation, into the architecture for crack information quickly extraction and model’s parameter reduction. We test latency and accuracy trade-offs of our proposed model using the DeepCrack Dataset. The results demonstrate strong performance in both accuracy and efficiency compared to other state-of-the-art models including DeepLabV3, FCN, LRASPP, Enet, Unet and DeepCrack. It is promising that ECSNet obtains the second place with an F1 score of (84.45%) and an Intersection over Union (IoU) of 73.08%. Furthermore, our model gains the largest Frames Per Second (FPS) and lowest training time among all the models which is 73.29 and 5011 seconds, respectively. It maintains a good balance between accuracy and efficiency metrics.

I. Introduction

Crack is a common defect of pavement that directly affects pavement service ability and driving safety [1]. Traditionally, the cracks are counted and measured by engineers, but it is time-consuming and labor-intensive [2]. With the development of image-based technologies, pavement crack detection approaches have achieved critical improvements in accuracy and reliability [3]. Crack classification and segmentation are two main focuses of the pavement crack detection. The purpose of classification is to differentiate different crack types while segmentation is a pixel-level extraction of crack from the background [4]. Segmenting cracks from pavement background is a more challenging task than classification and is more attractive to engineers and researchers. During the past decades, a series of crack segmentation methods have been proposed and deep learning-based techniques have become the most interesting and advanced approach to segmenting pavement cracks. This is because deep learning models learn from large-scale data and require little human involvement during training, which can dramatically increase their accuracy and robustness in segmentation tasks [5].