Bed Surface Profiling (BSP) data used for calculating pickup probability

Bed Surface Profiling (BSP) data

The Bed Surface Profiling (BSP) data scanned by an underwater 3D laser scanner for different flow conditions (Case Nos. 1-12). The sampling rate was 10 Hz. Each line of data corresponds to the xyz point cloud data, where x-direction corresponds to streamwise direction, y-direction to trnasverse dirction, and z-dirction to vertical direction. 

Code

A python code is included to process the BSP data for calculating sediment pickup probability for each case. This code analyzes 12 case files (Case1.txt to Case12.txt) containing measurement data, processes them using dynamic denoising filters, and calculates a derived metric (p_value) for each case. The results are saved in an Excel file.

Key Components

1. Dynamic Denoising Filter (dynamic_denoise_filter):

   • Implements two filtering schemes based on parameter d:
     • 3-Window Rule (d=0.00185): Checks immediate neighbors (previous and next columns) to correct 0, 1, or -1 values.
     • 11-Window Rule (d=0.00380): Uses a larger window (11 columns) with different criteria:
       • For zeros: Requires immediate neighbors (columns 4 and 6) to be zero.
       • For ±1: Requires all non-center columns (first 5 and last 5) to match the target value.

2. Case Processing (process_case):

   • Data Loading: Reads CSV-like files, extracts y (2nd column) and z (3rd column) values.
   • Derivative Calculation: Computes dz as the finite difference of z values scaled by d.
   • Marker Matrix: Creates a matrix m marking regions where dz exceeds ±0.
   • Denoising: Applies dynamic_denoise_filter to clean the marker matrix.
   • p-value Calculation:
     • Computes y_diff as half the absolute difference between adjacent y values.
     • Identifies valid transitions (1 → -1) in the denoised matrix.
     • Aggregates weighted counts and normalizes by the total y range.

3. Parameter Settings:

   • **d Values**:
     • 0.00185 for Cases 1–6
     • 0.00380 for Cases 7–12
   • Critical Value (cr): Set to 0 for thresholding dz.
   • Data Directory: Configured as D:/data_ana_ca.

4. Output:

   • Results are saved in analysis_result_ca_cr.xlsx with columns Case and p_value.

Usage

• Input: 12 text files (Case1.txt to Case12.txt) with comma-separated data.
• Output: Excel file containing calculated p-values.
• Runtime Monitoring: Prints processing time for each case during execution.

The code efficiently handles large datasets using vectorized operations and sliding window techniques, ensuring scalability for similar analytical tasks.