METHOD FOR OBJECT RECOGNITION USING QUEUE-BASED MODEL SELECTION AND OPTICAL FLOW IN AUTONOMOUS DRIVING ENVIRONMENT, RECORDING MEDIUM AND DEVICE FOR PERFORMING THE METHOD

摘要

The object recognition method using queuing-based model selection and optical flow in an autonomous driving environment uses a matrix-type density flow by calculating the optical flow of images continuously photographed in time by a sensor for an autonomous vehicle. Pre-processing the data; Generating a vectorized confidence threshold indicating a probability that a moving object exists for each cell of the preprocessed matrix to generate a confidence mask; Mapping the temporally continuously photographed image to the confidence mask to determine the presence or absence of a moving object on the image; And selecting an object recognition model using a tradeoff constant between the accuracy of object recognition and the stability of the queue for each time unit. Accordingly, by applying the optical flow to the confidence threshold of the object recognition system, it is possible to increase performance by applying the optical flow to object recognition in an autonomous driving environment.