Method and device for vehicle detection and tracking
A vehicle detection and vehicle technology, which is applied in the field of vehicle detection and tracking methods and devices, can solve the problems of reducing the accuracy of positioning the vehicle position, increasing the vehicle position detection error, etc.
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0100] see figure 2 , the present embodiment provides a vehicle detection and tracking method, comprising the following steps:
[0101] Step 200: Detect the currently captured vehicle image to obtain vehicle features of the vehicle image; wherein the vehicle feature includes at least one of the following: image grayscale feature, gradient magnitude feature, and gradient direction feature.
[0102] The vehicle image is each frame of the vehicle image captured by the driving recorder or the vehicle panoramic imaging system pre-installed on the vehicle through the camera of the driving recorder or the vehicle panoramic imaging system during the driving of the vehicle.
[0103] Image grayscale features, gradient magnitude features and gradient direction features can be obtained using the improved aggregation channel feature algorithm (SACF); of course, other existing image grayscale features and gradient magnitude features that can obtain vehicle images can also be used And the ...
Embodiment 2
[0180] see image 3 , this embodiment provides another vehicle detection and tracking method, the vehicle detection and tracking method can be divided into two main steps of detection and tracking, and video-based vehicle detection can be divided into two stages of training and detection. Video-based vehicle detection is trained by using machine learning methods. The features and feature calculation methods in the vehicle detection process are designed. Mixed channel features are used as vehicle features for training and detection. The trainer uses cascaded Adaboost Algorithm, the detection adopts the multi-scale multi-model detection mode combined with global and local, and the vehicle tracking adopts the detection correlation integration method to achieve fast vehicle tracking effect.
[0181] The feature calculation method provided in this embodiment is described as follows:
[0182] The vehicle detection features are described by SACF features, which include the combinati...
Embodiment 3
[0218] see Figure 4 , the present embodiment provides a vehicle detection and tracking device for performing the above vehicle detection and tracking method, including: a vehicle feature detection module 400, a model selection module 402, a position information determination module 404 and a vehicle tracking module 406;
[0219]The vehicle feature detection module 400 is used to detect the currently captured vehicle image to obtain the vehicle feature of the vehicle image; wherein the vehicle feature includes at least one of the following: image grayscale feature, gradient amplitude feature and gradient direction feature;
[0220] The model selection module 402 is connected with the vehicle feature detection module 400, and is used to select a group of models adapting to the vehicle image size from the pre-trained multi-model network structure;
[0221] The location information determination module 404 is connected with the model selection module 402, and is used to perform v...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More