Parking space detection method based on deep learning

Abstract

The invention discloses a parking space detection method based on deep learning, and the method comprises the steps: building a Faster R-CNN parking space detection model based on deep learning, training the parking space detection model, and enabling the Faster R-CNN parking space detection model to comprise a convolution layer, an RPN network layer and an RoI pooling layer; the convolutional layer is used for performing feature extraction to obtain a feature layer; the RPN network comprises a classification layer and a regression layer; a classification layer extracts a candidate area from the feature layer and judges whether the candidate area contains a target or not, and if the target exists, the regression layer is utilized to adjust the position and size of the candidate area; the RoI pooling layer maps the candidate region obtained by the RPN network layer as a candidate frame to the convolution layer, continues to classify and further adjusts the position and the size of the candidate region until the most accurate candidate region is obtained; and the Faster R-CNN parking space detection model is trained and tested, so that the establishment of the Faster R-CNN parking space detection model is completed. By using the constructed Faster R-CNN parking space detection model, high-accuracy detection of the parking space can be completed under the condition of complex image information.

Description

technical field [0001] The invention belongs to the technical field of intelligent driving and automatic parking, in particular to a parking space detection method based on deep learning Background technique [0002] With the improvement of national economic level, the total number of automobiles in our country is constantly increasing, which not only causes traffic congestion on the road to become normal, but also makes parking a difficult problem for drivers themselves. Therefore, intelligent driving came into being, especially the emergence of automatic parking technology to solve most of the parking problems. At present, automatic parking technology is mainly divided into two technologies: radar-based automatic parking and camera-based automatic parking. Among them, radar-based automatic parking mainly relies on sensors to measure the distance from surrounding objects, and then transmits them to the central processing unit. However, it is difficult to achieve accurate p...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the deficiencies in the prior art, the present invention proposes a parking space detection method based on deep learning. Using the constructed Faster R-CNN parking space detection model, it can complete high-level parking space detection under the condition of very complex image information. Accuracy detection

Images