December 2020
tl;dr: Parking slot detection by detecting marking point with a CenterNet-like algorithm.
Overall impression
For my future self: Dataset is super important. Your algorithm is only going to evolve to the level your dataset KPI requires it to.
The algorithm only focuses on detecting the marking point detection and did not mention too much about the post-processing needed to combine the marking points to parking slot. It is more general in that it can detect more than T/L-shaped marking points.
The paper is very poorly written, with tons of sloppy annotation and non-standard terminology.
Key ideas
- A coarse-to-fine marking point detection algorithm. Very much like CenterNet.
- The regression also predicts the “vertex paradigm”. Basically it predicts the pattern of the connectivity among the marking points.
'Vacant Parking Slot Detection in the Around View Image Based on Deep Learning.' Sensors 20.7 (2020): 2138. 'Vision-Based Parking Slot Detection Based on End-to-End Semantic Segmentation Training.' 2020 IEEE International Conference on Consumer Electronics (ICCE). 'End-to-End Trainable One-Stage Parking Slot Detection Integrating. The parking-slot detection module takes the surround-view image as the input, detects the parking-slots, and finally sends their physical positions with respect to the vehicle-centered coordinate system to the decision module for further process. Representative work on vision-based parking-slot detection will be reviewed as follows. Detects whether parking space is available or occupied using OpenCV image processing and computer vision tools. The algorithm only focuses on detecting the marking point detection and did not mention too much about the post-processing needed to combine the marking points to parking slot. It is more general in that it can detect more than T/L-shaped marking points. The paper is very poorly written, with tons of sloppy annotation and non-standard terminology. Firstly, to facilitate the study of vision-based parking-slot detection, a large-scale parking-slot image database is established. For each image in this database, the marking-points and parking-slots are carefully labeled. Such a database can serve as a benchmark to design and validate parking-slot detection algorithms.
Technical details
- Annotated a dataset (~15k images). This is slightly bigger than PS2.0 dataset with 12k images.
- The paper uses L2 loss to supervise the heatmaps and attributes. This is a bit strange as most studies uses focal loss for heatmap prediction and L1 for attribute prediction.
Notes
Parking Slot Detection Github Extension
- Questions and notes on how to improve/revise the current work