ECCNet: Efficient chained centre network for real‐time multi‐category vehicle tracking and vehicle speed estimation

Abstract

To achieve the intelligent perception of traffic video and overcome the dispersion of computing modules and the separation processing of multiple tasks, efficient chained centre network (ECCNet) is proposed as a unified framework, accomplishing detection, vehicle classification, tracking, and vehicle speed estimation simultaneously. First, for the speedaccuracy trade-off, CA-CenterNet is presented, which can detect vehicles and classify vehicle types to serve cross-frame tasks more accurately by embedding coordinate attention. Second, 3D convolution is employed to construct a self-adaptive branch for data association and speed estimation, respectively. This self-adaptive approach leverages the power of deep learning to enhance tracking performance and avoid camera calibration via capturing motion information across frames. Moreover, a chained structure is adopted to reuse the backbone feature map. The spatio-temporal information of adjacent frames can be extracted at almost no additional cost. Finally, the above single-frame and cross-frame tasks are integrated into a unified multi-task collaborative optimization model. The effectiveness of ECCNet is verified with experiments on the UA-DETRAC dataset. ECCNet achieves 55.5% MOTA, 0.76 F1 value, and 3.10 MAE with an inference speed of 32.5 Hz in the tracking, classification, and speed estimation tasks, respectively.

Instant-level vehicle speed and traffic density estimation using deep neural network

“…Figure 6 shows images selected from the output of this methodology across the four scenes considered i.e., sunny, cloudy, rainy. Our approach outperforms [29] across all scenes.…”

“…The model's input are videos and the vehicle's maximum speed in each video footage for training and the predicted speed is a function of the learned local movement and the maximum speed. However, techniques have been proposed which are invariant to some of these assumptions such as the efficient chained centre network (ECCNet) [29], which is a unified framework for accomplishing vehicle detection, tracking speed estimation in parallel. This paper follows a simple mathematical approach invariant to road length and maximum speed limit of the road, where the pixel-per-second distance traveled by a tracked vehicle is computed and fed into a regression model trained on the UA-DETRAC dataset, and the regression model in training intuitively learns the weights required to map pixels to meters and make instant level speed predictions.…”

Desenvolvimento de um sistema de detecção e rastreamentode veículos para análise de anomalias de tráfego em rodovias utilizando estruturas espaciais e temporais por meio de Visão Computacional

“…ECCNet: Efficient chained centre network for real-time multi-category vehicle tracking and vehicle speed estimation (ECCNet) (YU et al, 2022) é um um avançado algoritmo de rastreamento em tempo real de veículos multi-categoria e estimativa de velocidade. O ECCNet utiliza uma rede neural profunda eficiente e é composto por três módulos principais: detecção, rastreamento e estimativa de velocidade.…”