Robust Parking Block Segmentation from a Surveillance Camera Perspective

Abstract

Parking block regions host dangerous behaviors that can be detected from a surveillance camera perspective. However, these regions are often occluded, subject to ground bumpiness or steep slopes, and thus they are hard to segment. Firstly, the paper proposes a pyramidal solution that takes advantage of satellite views of the same scene, based on a deep Convolutional Neural Network (CNN). Training a CNN from the surveillance camera perspective is rather impossible due to the combinatory explosion generated by multiple point-of-views. However, CNNs showed great promise on previous works over satellite images. Secondly, even though there are many datasets for occupancy detection in parking lots, none of them were designed to tackle the parking block segmentation problem directly. Given the lack of a suitable dataset, we also propose APKLOT, a dataset of roughly 7000 polygons for segmenting parking blocks from the satellite perspective and from the camera perspective. Moreover, our method achieves more than 50% intersection over union (IoU) in all the testing sets, that is, at both the satellite view and the camera view.

[Retracted] Automatic Parking System Based on Improved Neural Network Algorithm and Intelligent Image Analysis

“…where O is the path index set of traversing obstacles and ζ μ is the transmission loss. ere are two uncertain factors in formula (5), which affect the accuracy of the detection signal. One is the interference of adjacent paths, and the other is that the path delay time is not an integral multiple of the sampling time T s , which is missed in the sampling.…”

“…The existing indoor positioning models include BP neural network positioning model, WLAN positioning model, support vector machine model, ZigBee fuzzy clustering positioning model, etc. [ 5 ]. In this study, WiFi coverage is used to identify and locate the free parking spaces.…”

Desafios na Gerência de Estacionamentos por Imagem

“…Métodos para a segmentação automática de vagas de estacionamento utilizando técnicas clássicas de processamento de imagens, como transformações de perspectivas, binarizações e detectores de borda, são propostos em [Bohush et al 2018, Zhang et al 2019, Zhang and Du 2020. Extratores de características são usados em [Vítek and Melničuk 2018] para detectar se blocos da imagem pertencem ou não a carros, e métodos baseados em redes neurais convolucionais são apresentados em [Agrawal and Urolagin 2020, Padmasiri et al 2020, Hurst-Tarrab et al 2020.…”

“…Apesar de existirem métodos tratando da segmentação automática de vagas, os trabalhos são escassos e em sua grande maioria apresentam dados insuficientes sobre os métodos, experimentos e principalmente resultados. Dos trabalhos citados, apenas [Padmasiri et al 2020, Hurst-Tarrab et al 2020] apresentam resultados quantitativos.…”

“…Issoé um desafio, já que não se pode esperar que tenhamos coordenadas precisas de latitude e longitude de uma quantidade razoável de pontos das imagens das câmeras. Dos trabalhos pesquisados, apenas [Hurst-Tarrab et al 2020] tenta realizar a correspondência entre imagens de satélite e as imagens de câmeras de vigilância de estacionamentos. Isso indica que pesquisas nesse tópico ainda são necessárias.…”