Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Deep learning-based haze visibility detection method

A deep learning and detection method technology, applied in the field of image processing, can solve the problems of expensive equipment, not particularly convenient to use, no model, etc., to achieve the effect of rapid classification and recognition

Inactive Publication Date: 2017-10-20
NANJING UNIV OF POSTS & TELECOMM
View PDF1 Cites 22 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This often brings a lot of human errors, and the instrument measurement method refers to the detection of the visibility value through some optical components. This kind of equipment is usually expensive and not particularly convenient to use. Therefore, in recent years, there have been more Many people pay attention to the image-based haze visibility detection, mainly through the calculation of the visibility value through the estimation of the atmospheric extinction coefficient, which is a complex calculation process
Based on the above, some people began to consider using machine learning methods to identify visibility, but they are only assumptions, and there is no specific model for the time being

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
  • Deep learning-based haze visibility detection method
  • Deep learning-based haze visibility detection method
  • Deep learning-based haze visibility detection method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0021] The technical scheme of the present invention will be further described in detail below in conjunction with the accompanying drawings:

[0022] The present invention is specifically a haze visibility detection algorithm that uses a self-built traffic image database and uses a convolutional neural network as a training and verification network for the algorithm. The specific implementation process will be described in detail below.

[0023] Such as figure 1 Shown is the flow chart of the present invention. The implementation of the method for detecting the visibility of haze based on deep learning of the present invention mainly includes the following steps:

[0024] Step 1: Create a road traffic haze image library

[0025] In terms of data sets, two data sets will be used. The first data set is used for training and the second data set is used for verification. According to experience, an 8:2 ratio is used for data distribution. According to Article 81 of the "Implementation Re...

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

The invention discloses a deep learning-based haze visibility detection method. The method comprises the following steps of: firstly creating a road traffic haze image library which comprises a training sample set and a cross validation sample set; preprocessing all the haze images in the traffic haze image library; extracting farthest visibility edge features of the preprocessed haze images I the training sample set by adoption of a convolutional neural network so as to obtain a plurality of feature maps; forwardly propagating the feature maps to a configured convolutional neural network to be trained, adjusting weights between layers in the convolutional neural network through a counter-propagation algorithm, carrying out repeated iteration to solve a convolutional neural network model for haze image classification, and optimizing the convolutional neural network model through the preprocessed cross validation sample set so as to finally obtain a visibility detection model for haze image classification; and carrying out classification judgement on pictures shot by a pavement by utilizing the obtained visibility detection model so as to detect the haze condition in real time.

Description

Technical field [0001] The invention relates to the technical field of image processing, in particular to a method for detecting visibility of haze based on deep learning. Background technique [0002] In recent years, haze weather has occurred frequently, and meteorological experts have analyzed many causes of haze weather. It is reported that the continuous haze weather has brought more serious impacts on all walks of life. The first is the impact on transportation. Local flights affected by the haze have been delayed or even cancelled, the high-speed public facilities have been closed, and maritime traffic has also suffered. Different degrees of influence. Including railways, “fog flashes” caused by smog have caused power outages, temporary stops or delays on the Beijing-Guangzhou Railway. Accurate visibility detection is one of the necessary links to alleviate this problem. [0003] Traditional smog visibility detection methods are divided into human eye visual inspection met...

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
IPC IPC(8): G06T7/00G06K9/46G06K9/62G06N3/04
CPCG06T7/0002G06T2207/30192G06V10/44G06V10/60G06N3/045G06F18/214
Inventor 许艳丽李海波成孝刚邵文泽吕泓君
Owner NANJING UNIV OF POSTS & TELECOMM
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
Eureka Blog
Learn More
PatSnap group products