Record a Lap Using a MATLAB Script

Aus HSHL Mechatronik
Zur Navigation springen Zur Suche springen

record, storing, data manipulation, calibration

Record a Lap

After successfully recording the lab, an essential step is to provide quality data for the AI by cleaning and filtering the data(removing noice). These can be archieved through the following.

data manipulation

  • Normalization:

is the set of steps applied to raw data to make it clean and suitable for analysis or modelling. It covers tasks like handling missing values, encoding categories, removing outliers, and feature scaling.

Tabelle 5: Matlab Setup For Normalisation
# Name Description Pictures
1 Standardization: Ensuring every image treated is of the same size and dimension
Fig. 5: Camera setup for image with calibration


  • Pre-processing:
Tabelle 6: Matlab Setup for Pre-processing
# Name Description Pictures
1 handling missing values: Check for valid and delete missing steering or frames
Fig. 5: Camera setup for image with calibration
2 removing outliers: Filter and remove unnecessary layer before training
Fig. 5: Camera setup for image with calibration

calibration

There are unusual distortions caused by the camera lens on the image; these distortions can cause noises in the input, hence reducing the AI's accuracy. Fixing these distortions can be done through calibration of the camera [1]. screenshot

Results

Tabelle 6: Calibration Output
# Name Description Pictures
1 Results:
Fig. 5: Camera setup for image without calibration
Fig. 5: Camera setup for image with calibration