Railway Track Weeding Sensor Fusion for Precise Herbicide Targeting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing weeding apparatuses on railway vehicles lack optimal data fusion methods, leading to imprecise or redundant weed-killing interventions due to inadequate consideration of data quality affected by external conditions, resulting in waste of weed-killing agents or missed weed species.
Innovation Solution
A weeding apparatus and method that utilizes a combination of visible spectrum and infrared cameras, LIDAR, and a processing module with convolution operations and multiplicative factors based on weather and brightness conditions to enhance data reliability and precision in weed detection, ensuring targeted application of weed-killing agents.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If data arrays from multiple sensors are re-processed through convolution procedures without using different multiplicative factors, then the processing is simpler, but the data fusion is not optimum leading to imprecise or redundant interventions
Solution Approach 1:
The patent applies local quality by assigning different multiplicative factors to different sensor data arrays based on their individual qualities and external conditions. Each sensor's data is weighted differently in the convolution operation, allowing the system to optimize detection precision by emphasizing more reliable data sources under specific conditions while maintaining a unified processing framework.
2Loss of substance
If uniform multiplicative factors are used for all sensor arrays, then the processing method is simpler, but waste of weed-killing agent occurs due to imprecise interventions
Solution Approach 1:
The patent implements parameter changes by dynamically adjusting multiplicative factors based on external conditions such as weather, lighting, and sensor performance. This allows the system to optimize the weighting of each sensor's data in real-time, improving intervention precision and reducing weed-killing agent waste by avoiding both false positives and false negatives.
3Reliability
If multiple sensors with different detection capabilities are used, then the detection coverage is improved, but the data fusion becomes suboptimal without quality-based weighting
Solution Approach 1:
The patent applies local quality by assigning different multiplicative factors to different sensor data arrays based on their individual qualities and external conditions. Each sensor's data is weighted differently in the convolution operation, allowing the system to optimize detection precision by emphasizing more reliable data sources under specific conditions while maintaining a unified processing framework.
4Productivity
If data arrays are processed without considering external conditions affecting data quality, then the processing is faster, but redundant or missed interventions occur
Solution Approach 1:
The patent implements preliminary action by pre-defining quality assessment criteria and multiplicative factor ranges for different sensor types and external conditions. This allows the system to quickly apply appropriate weights during real-time processing without complex on-the-fly calculations, maintaining high processing speed while improving detection accuracy through condition-aware data fusion.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enhances the accuracy and efficiency of weed detection and treatment by weighing data sources based on reliability, reducing waste and improving operational precision.
Implementation Method 1
a first camera (3) which is configured to detect at least the visible spectrum
Implementation Method 2
a second camera (5) which is configured to detect at least the infrared spectrum
Implementation Method 3
a remote-sensing device, of the LIDAR type, comprising a laser emitter
Data Source
Figure 1
Figure 2~3
Figure 4a~4b
AI summary
A weeding apparatus is arranged to be installed on a railway vehicle for weeding, and comprises at least two sources of data, a processing module and a control module whereon a first and a second software are implemented, respectively, the apparatus being arranged through said software to process said data and to combine them according to "customized" multiplicative factors based upon the atmospheric conditions and the alternation of day and night.