SVM classifier, method and apparatus for discriminating vehicle image therewith

A classifier and vehicle technology, applied in the field of vehicle image recognition, can solve problems such as large number of support vectors, poor real-time performance, poor recognition effect, etc., achieve good adaptability, increase the overall recognition rate, and improve real-time performance

Inactive Publication Date: 2011-10-26
NEUSOFT CORP +1
View PDF1 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] However, for vehicles on the actual driving road, it not only includes the requirement to identify vehicles of various types, various colors, distances from the vehicle, and different angles from the direction of the vehicle, but also requires the identification of different backgrounds, different lighting and different weather conditions. Therefore, only a single SVM classifier was used for classification and recognition in the past. On the one hand, it will lead to a large training sample set, and it is difficult for a single SVM classifier to distinguish samples with complex distributions (such as using a single SVM classifier When identifying light and dark mixed vehicles, the darker vehicle recognition effect is not good), so the recognition rate is not high. Due to the long time of recognition processing, the real-time performance is not good

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • SVM classifier, method and apparatus for discriminating vehicle image therewith
  • SVM classifier, method and apparatus for discriminating vehicle image therewith
  • SVM classifier, method and apparatus for discriminating vehicle image therewith

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0038] Embodiment 1 uses the characteristics of the horizontal edge and the vertical edge of the image to train the corresponding classifier.

[0039] In vehicle image recognition, the vehicle or background in the image to be recognized has several horizontal and vertical edges (horizontal and vertical lines).

[0040] The number of horizontal lines is judged for vehicle samples and background samples, and the method for extracting horizontal line features can be difference method, edge operator and other methods. The present invention uses the difference method to extract the horizontal line feature as an example, and the description will be made in detail.

[0041] First, convert the sample image to a grayscale image. Assuming that the sample grayscale image to be processed has a total of n rows, for the convenience of description, each row is numbered from "1" from the bottom to the top, and then all the rows from the second to the nth Row subarea, such as Figure 7 Shown. Every...

Embodiment 2

[0052] In Embodiment 2, the characteristics of brightness, contrast, and color of the image are used to train the corresponding classifiers.

[0053] First, perform light and dark judgment (brightness judgment) on the training sample image. The available methods include judgment based on the average gray value of the entire image or the gray value of the pixels in the image. The present invention uses the gray level of the pixels in the image. Value for judgment, such as Figure 8 As shown, the training sample image is first converted into a gray image, and then the statistical gray value is less than the given value d 8 (For example, set 60≤d 8 ≤100) the number of pixels, if the percentage of the number of pixels meeting the condition in the number of pixels in the entire image is greater than the given value d 9 (E.g. 0.5 9 ≤0.8), the training sample image is determined to be a low-luminance sample; otherwise, it is determined to be a high-luminance sample. Finally, use a set o...

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

No PUM Login to view more

Abstract

A method of utilizing SVM grader to identify vehicle image includes classifying train sample as per vehicle identification character, obtaining various SVM graders corresponding to multiple character by training SVM graders separately, picking up characters of image to be identified and distributing image to be identified into SUM grader corresponding to the character for carrying out identification on vehicle image when vehicle image identification is carried out.

Description

Technical field [0001] The present invention relates to the technology of vehicle image recognition, in particular to the SVM (Support Vector Machines) classifier in vehicle image recognition, and a method and device for recognizing vehicle images by using the SVM classifier. Background technique [0002] The SVM (Support Vector Machines) classifier in vehicle image recognition is a general machine learning method based on the framework of statistical learning theory. It was originally proposed for two types of classification problems. It has a simple structure and generalization ability. Strong advantages. [0003] In vehicle image recognition, when the SVM classifier is used for vehicle image recognition, the past methods are as follows figure 1 As shown, that is, firstly use manually selected vehicle training samples and background training samples for training in the training process to obtain an SVM classifier, and then use various features (such as vehicle bottom shadow, hori...

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
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/62
Inventor 文学志孙雷
Owner NEUSOFT CORP
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products