Generalized Parking Occupancy Analysis Based on Dilated Convolutional Neural Network

Abstract

The importance of vacant parking space detection systems is increasing dramatically as the avoidance of traffic congestion and the time-consuming process of searching an empty parking space is a crucial problem for drivers in urban centers. However, the existing parking space occupancy detection systems are either hardware expensive or not well-generalized for varying images captured from different camera views. As a solution, we take advantage of an affordable visual detection method that is made possible by the fact that camera monitoring is already available in the majority of parking areas. However, the current problem is a challenging vision task because of outdoor lighting variation, perspective distortion, occlusions, different camera viewpoints, and the changes due to the various seasons of the year. To overcome these obstacles, we propose an approach based on Dilated Convolutional Neural Network specifically designed for detecting parking space occupancy in a parking lot, given only an image of a single parking spot as input. To evaluate our method and allow its comparison with previous strategies, we trained and tested it on well-known publicly available datasets, PKLot and CNRPark + EXT. In these datasets, the parking lot images are already labeled, and therefore, we did not need to label them manually. The proposed method shows more reliability than prior works especially when we test it on a completely different subset of images. Considering that in previous studies the performance of the methods was compared with well-known architecture—AlexNet, which shows a highly promising achievement, we also assessed our model in comparison with AlexNet. Our investigations showed that, in comparison with previous approaches, for the task of classifying given parking spaces as vacant or occupied, the proposed approach is more robust, stable, and well-generalized for unseen images captured from completely different camera viewpoints, which has strong indications that it would generalize effectively to other parking lots.

Video-Based Parking Occupancy Detection for Smart Control System

“…AlexNet 96.54% ResNet50 96.24% CarNet [21] 97.24% Ours 98.97% For testing purposes, the test subset of the CNRPark + EXT dataset was used. The accuracy is used to evaluate the performance of the identification task.…”

“…For parking detection, Acharya and Khoshalham [20] proposed a system for real-life situations by providing functionalities for real-time monitoring, and by modifying a convolutional neural network (CNN) to classify whether the space is occupied or free. Nurullayev and Lee [21] proposed methods based on a CNN designed to detect parking grid occupancy in parking lots by using an open dataset to evaluate a detection model. Object detection can produce the same effects, and many methods apply object detection in parking sections, such as Faster region CNN (R-CNN) [22], single-shot detection [23], and You Only Look Once (YOLO) [24].…”

Desafios na Gerência de Estacionamentos por Imagem

“…A classificação das vagas pode ser caracterizada como o processo de se inferir se determinada imagem (previamente recortada como na Figura 1) contendo exatamente uma vaga de estacionamento encontra-se ocupada ou vazia. Esteé um problema amplamente estudado, sendo que os métodos propostos comumente se baseiam na extração de características para utilização posterior de classificadores [de Almeida et al 2015, Ahrnbom et al 2016, Bohush et al 2018, Vítek and Melničuk 2018, Varghese and Sreelekha 2019, ou na utilização de redes neurais convolucionais [Amato et al 2017, Nurullayev and Lee 2019, Jensen et al 2017, Nieto et al 2018, Zhang et al 2019, Chen et al 2020 Para os testes dos métodos de classificação de vagas individuais, podemos definir os seguintes cenários:…”

Vacant Parking Slot Detection in the Around View Image Based on Deep Learning

“…The mAlexNet was inspired by the AlexNet, where the number of filters and neurons was reduced to improve real-time performance. Nurullayev and Lee [41] designed a generalized parking occupancy classification method for varying images captured from different camera views based on the dilated convolutional neural network. Considering that parking slot occupancy classification is a simple two-category task, a small number of dilated convolutional layers and large kernel sizes were utilized to avoid learning with too deep models.…”