A long-distance traffic sign detection and recognition method suitable for in-vehicle systems
A traffic sign and vehicle-mounted system technology, which is applied in the field of long-distance traffic sign detection and recognition, can solve the problem of low precision and achieve the effects of reducing hardware storage, saving computing resources, increasing security and device response time
Active Publication Date: 2022-05-06
NORTHEAST GASOLINEEUM UNIV
View PDF8 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0004] The purpose of the present invention is to provide a long-distance traffic sign detection and recognition method suitable for vehicle-mounted systems, which is used to solve the problem of low precision in the long-distance traffic sign detection and recognition of existing short-distance detection and recognition methods
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment
[0176] The resolution of each traffic scene in the model training dataset Tsinghua-Tencent 100K is 2048×2048 pixels, and the size of traffic signs is between 0-32 pixels and 32-96 pixels, accounting for 41.6% and 49.1% of the data set respectively , that is, the size of 90.7% of the traffic signs accounts for less than 1% of the traffic scene, which belongs to the long-distance traffic sign detection and recognition situation.
[0177] (1) Data set processing: the FL-CNN model adopted in the present invention is in the training process, in order to keep the balance of the sample set, the method of resampling is taken for the category of traffic scenes where each type of traffic sign is less than 100. The ratio of training set and test set is 1:2;
[0178] (2) Contrast index: the present invention adopts the measurement index F1-measure commonly used to measure the accuracy rate as the detection and identification index in the specific testing process, and the larger the value ...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention relates to a long-distance traffic sign detection and recognition method suitable for vehicle-mounted systems, which includes: 1. Preprocessing the traffic sign image sample set; 2. Constructing a lightweight convolutional neural network to complete the identification of traffic signs Convolutional feature extraction; 3. Construct the attention feature map through the channel-spatial attention module embedded in the lightweight convolutional neural network; 4. Use the region generation network RPN to generate candidate regions for the target; 5. Generate RPN 6. Send the feature vector to the fully connected layer to output the category and position of the traffic sign; 7. Establish an attention loss function to train the FL‑CNN model; 8. Repeat 2 to 7 to complete the sample training of the FL-CNN model; 9. Repeat 2 to 6 to complete the traffic sign detection and recognition in the actual scene. The invention realizes the detection and recognition of long-distance traffic signs, and the accuracy reaches 92%.
Description
1. Technical field: [0001] The invention relates to the field of intelligent transportation for unmanned driving and assisted driving, and solves the long-distance detection and recognition method of road traffic signs, and specifically relates to a long-distance traffic sign detection and recognition method suitable for vehicle-mounted systems. 2. Background technology: [0002] In the field of intelligent transportation, traffic sign detection and recognition is an important research issue for systems such as unmanned driving and assisted driving. A lot of research work has been done on this at home and abroad, but there are still great deficiencies, which cannot be practically applied to practice. The reasons are as follows: (1) Traditional detection and recognition methods designed based on features such as color and shape are less robust in the face of sign deformation, motion blur, weather, etc. in actual traffic scenes, and are difficult to implement in practice. (2)...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06V20/58G06V10/774G06V10/82G06K9/62G06N3/04
CPCG06V20/582G06N3/045G06F18/214
Inventor 刘志刚杜娟田枫韩玉祥高雅田张可佳
Owner NORTHEAST GASOLINEEUM UNIV