Target vehicle detection method based on deep learning

Abstract

The invention relates to a target vehicle detection method based on deep learning, and the method comprises the following steps: 1) collecting tail feature point cloud data of a target vehicle throughtwo single-line laser radars disposed at the tail of a parking robot, and carrying out the preprocessing of the tail feature point cloud data, and obtaining a binary image; 2) marking the binary image, obtaining the position of the tail of the target vehicle, and generating a training data set; 3) constructing a deep convolutional neural network suitable for target vehicle detection and a loss function of the deep convolutional neural network; and 4) augmenting the training data set, inputting the augmented training data set into the deep convolutional neural network, carrying out training updating on parameters in the convolutional neural network according to the difference between the output value and the training true value to obtain optimal network parameters, and carrying out detection according to the trained deep convolutional neural network. Compared with the prior art, the method has the advantages of high robustness, no dependence on manual characteristics, low detection cost and the like.

Description

Technical field [0001] The invention relates to the technical field of intelligent parking, in particular to a target vehicle detection method based on deep learning. Background technique [0002] In the field of intelligent driving, the detection of target vehicles is one of the key tasks to ensure the safe driving of unmanned vehicles. Also in the field of intelligent parking technology, detecting the position of the target vehicle is a key step to achieve the precise alignment of the parking robot to the target vehicle. Since lidar is less affected by the environment and can collect accurate point cloud data of the target vehicle, lidar has become the most important vehicle detection and positioning sensor in the field of smart parking. [0003] At present, in the field of intelligent parking technology, the detection method of the target vehicle mainly uses the traditional detection algorithm based on the manual feature of the target vehicle. Although the traditional detectio...

AI Technical Summary

Problems solved by technology

Although the traditional detection algorithm has a small amount of calculation and a fast speed, in many scenarios, the features of the target vehicle cannot match the manual features well, resulting in low recall rate and poor robustness of the traditional target vehicle detection algorithm.

Images