Abstract:Due to the complex visual environment, such as lighting variations, shadows, and limitations of vision, the accuracy of vacant parking slot detection for the park assist system (PAS) with a standalone around view monitor (AVM) needs to be improved. To address this problem, we propose a vacant parking slot detection method based on deep learning, namely VPS-Net. VPS-Net converts the vacant parking slot detection into a two-step problem, including parking slot detection and occupancy classification. In the parking slot detection stage, we propose a parking slot detection method based on YOLOv3, which combines the classification of the parking slot with the localization of marking points so that various parking slots can be directly inferred using geometric cues. In the occupancy classification stage, we design a customized network whose size of convolution kernel and number of layers are adjusted according to the characteristics of the parking slot. Experiments show that VPS-Net can detect various vacant parking slots with a precision rate of 99.63% and a recall rate of 99.31% in the ps2.0 dataset, and has a satisfying generalizability in the PSV dataset. By introducing a multi-object detection network and a classification network, VPS-Net can detect various vacant parking slots robustly.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is that in complex visual environments (such as illumination changes, shadows, and limited fields of view), the accuracy of idle parking space detection by the Parking Assist System (PAS) relying solely on the Around - View Monitoring system (AVM) needs to be improved. Specifically, the paper proposes a deep - learning - based idle parking space detection method, VPS - Net, aiming to solve the following problems: 1. **Robustness in complex environments**: Traditional methods perform poorly in complex visual environments such as illumination changes and shadows, resulting in a decline in detection accuracy. VPS - Net improves robustness in these complex environments by combining deep - learning techniques. 2. **Detection of multiple types of parking spaces**: Different types of parking spaces (such as vertical parking spaces, parallel parking spaces, and oblique parking spaces) need to be effectively identified. VPS - Net can detect various types of parking spaces and can infer the complete parking space based on geometric cues. 3. **Real - time requirements**: Traditional parking space detection methods often involve complex post - processing steps and are difficult to meet real - time requirements. VPS - Net simplifies the detection process to achieve efficient real - time detection. 4. **Classification of parking space occupancy status**: It is not only necessary to detect the location of the parking space but also to determine whether it is idle. VPS - Net divides parking space detection into two stages: parking space detection and occupancy status classification, thereby more accurately identifying idle parking spaces. ### Specific solutions of VPS - Net VPS - Net decomposes the idle parking space detection problem into two steps: 1. **Parking space detection**: Use a YOLOv3 - based detector to simultaneously detect parking space markers and parking space heads. Match paired markers through geometric cues and infer the complete parking space location. 2. **Occupancy status classification**: Design a customized Convolutional Neural Network (DCNN) to distinguish whether the parking space is idle. To improve classification performance, the parking space image is normalized to a uniform size (120×46 pixels) and pre - processed by perspective transformation technology. Through these two steps, VPS - Net can efficiently and accurately detect and classify idle parking spaces in complex environments. Experimental results show that VPS - Net has a precision of 99.63% and a recall rate of 99.31% on the ps2.0 dataset, and achieves a real - time detection speed of 20.5 milliseconds per frame on the Nvidia Titan Xp.

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

Vacant Parking Slot Recognition Method for Practical Autonomous Valet Parking System Using around View Image

Holistic Parking Slot Detection with Polygon-Shaped Representations

Semantic segmentation-based parking space detection with standalone around view monitoring system

Visual Detection and Image Processing of Parking Space Based on Deep Learning

Review of Vision-Based Deep Learning Parking Slot Detection on Surround View Images

Automatic Vision-Based Parking Slot Detection and Occupancy Classification

A Universal Vacant Parking Slot Recognition System Using Sensors Mounted on Off-the-Shelf Vehicles

Parking Spot Classification based on surround view camera system

Semantic Segmentation of Panoramic Images for Real-Time Parking Slot Detection

VH-HFCN Based Parking Slot and Lane Markings Segmentation on Panoramic Surround View

PSDet: Efficient and Universal Parking Slot Detection

Deep-Learning-Based Parking Area and Collision Risk Area Detection Using AVM in Autonomous Parking Situation

Deep learning-based parking occupancy detection framework using ResNet and VGG-16

EPSDNet: Efficient Campus Parking Space Detection via Convolutional Neural Networks and Vehicle Image Recognition for Intelligent Human–Computer Interactions

Real Time Detection Algorithm of Parking Slot Based on Deep Learning and Fisheye Image

Parking Lot Occupancy Detection with Improved MobileNetV3

Global Perception-based Robust Parking Space Detection Using a Low-cost Camera

Predicting vacant parking space availability zone-wisely: a hybrid deep learning approach

Visual Parking Occupancy Detection Using Extended Contextual Image Information via a Multi-Branch Output ConvNeXt Network