Vacant Parking Space Detection based on TaskConsistency and Reinforcement Learning

Abstract

In this paper, we proposed a novel task-consistency learning method that allows training a vacant space detection network (target task) based on the logistic consistency with the semantic outcomes from a naive flow-based motion behavior classifier (source task) in a parking lot. By well designing the reward mechanism upon semantic consistency, we show the possibility to train the target network in a reinforcement learning setting. Compared with conventional supervised detection methods, the major contribution of this work is to learn a vacant space detector via semantic consistency rather than supervised labels. The dynamic learning property may make the proposed detector been deployed in different lots easily without heavy training loads. The experiments show that based on the task consistency rewards from the motion behavior classifier, the vacant space detector can be trained successfully.

Experiment

Several demos had been prepared to compare our RL based model with a supervised learning-based model. The first scenario is a light changing scenario. Other demos present performance under different car density levels. The sparse scenario means only a few cars in a parking lot, the for less occlusion pattern could be observed. The dense scenario means the parking lot is nearly full and all the car is affected by occlusion patterns.

Light Changing Scenario

Light Changing Scenario-Supervised learning with 34688 samples

Light Changing Scenario-RL learning with 824 training trajectories

Sparse Scenario

Sparse Scenario-Supervised learning with 34688 samples

Sparse Scenario-RL learning with 824 training trajectories

Moderate Scenario

Moderate Scenario-Supervised learning with 34688 samples

Moderate Scenario-RL learning with 824 training trajectories

Dense Scenario

Dense Scenario-Supervised learning with 34688 samples

Dense Scenario-RL learning with 824 training trajectories