Weather Recognition Using Multi-Kernel Classifier for Image Feature Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image feature extraction methods fail to account for weather conditions, leading to inaccurate feature extraction and subsequent failures in computer vision applications, as the same object captured in different weather conditions produces distinct image features.
Innovation Solution
A weather recognition method and device that utilizes a multi-kernel classifier to identify the weather by extracting specific image features using preset algorithms for clear, rainy, snowy, and smoggy days, incorporating image contrast and saturation analysis, and employing techniques like guided-filtering and HOG feature extraction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a same image feature extraction method is applied to images taken in different weather conditions, then the extraction process is simple and unified, but the image feature extraction accuracy deteriorates
Solution Approach 1:
The patent segments the image feature extraction process by creating different extraction algorithms tailored to specific weather conditions (clear weather, rainy weather, snowy weather, smoggy weather). Each weather type has its own dedicated extraction method that optimizes for that condition, thereby resolving the contradiction between unified simplicity and condition-specific accuracy.
Solution Approach 2:
The patent changes the parameters and methods of image feature extraction based on weather conditions. By detecting the weather type first, the system selects appropriate extraction parameters and algorithms for each weather scenario, improving extraction accuracy without requiring a completely complex unified system.
2Measurement precision
If weather-specific image feature extraction algorithms are used for each weather type, then the image feature extraction accuracy is improved, but the system complexity increases
Solution Approach 1:
The patent applies preliminary action by first detecting the weather condition before performing image feature extraction. This preliminary weather detection step allows the system to select the appropriate extraction algorithm in advance, avoiding the need for a single complex algorithm that would have to handle all weather conditions simultaneously.
Solution Approach 2:
The patent introduces an intermediary weather detection module that bridges the gap between the image input and the feature extraction process. This intermediary component identifies the weather type and directs the image to the appropriate extraction algorithm, simplifying the overall system architecture while maintaining high extraction accuracy.
3Reliability
If weather detection is performed before image feature extraction, then the subsequent image processing reliability is improved, but the processing time increases
Solution Approach 1:
The patent applies partial action by performing weather detection on only the critical regions or key features of the image that are most indicative of weather conditions, rather than analyzing the entire image in detail. This reduces the time overhead of weather detection while still providing sufficient accuracy for selecting the appropriate feature extraction algorithm.
Data Source
AI summary
The invention relates to a weather recognition method and device based on image information detection, including: obtaining an image extracting multiple first image features of the image with respect to each preset type of weather using a number of first preset algorithms preset correspondingly for different preset types of weather; inputting the multiple first image features to a preset multi-kernel classifier, the multi-kernel classifier performing classification according to the image features to identify the weather in which the image was taken. The multi-kernel classifier is realized by: selecting a first preset number of image samples for each of the preset types of weather; for the image samples of this type of weather, extracting the first image features of each image sample according to the first preset algorithm corresponding to this preset type of weather; and performing machine learning for the first image features according to a preset multi-kernel learning algorithm.