Pattern matching and Computer Vision

Each folder represents an individual program and has been appropriately named to identify its application. Note that the program for Moth Eye classification has been written in Python using OpenCV. Other programs have been written in Python and Numpy, Matplotlib, Scipy (without OpenCV or other CV libraries).

Required Packages

• Python 3.7
• numpy
• scipy
• OpenCV
• matplotlib
• PIL

Compilation and Execution Instructions

Data

The python files read data from a hardcoded path to the local folder.

Cross correlation in 1D and 2D

To excecute 1D spatial cross correlation algortithm, follow these steps:

1. cd exp1_signalMatching
2. Run python signal_offset_temporal.py to process files sensorData1.txt and sensorData2.txt
3. To toggle SSD, set use_SSD = True on line 13.
4. To toggle the library function, set use_library = True on line 14.
5. To toggle the handmade cross correlation, set use_convolution = True on line 15. Currently this method takes approximately one hour.
6. To toggle the faster handmade convolution method, set speed_up = True on line 16.
   

To excecute 1D spectral cross correlation algortithm, run python signal_offset_FT.py to process files sensorData1 and sensorData2.

To excecute 2D spatial cross correlation algortithm, follow these steps:

1. cd imageMatching
2. Run python spatial_2D_image_match.py to process files wally_puzzle_rocketman.png and wallypuzzle_png.png.
   

To excecute 2D spectral cross correlation algortithm, run python spectral_2D_image_match.py to process files wally_puzzle_rocketman.png and wallypuzzle_png.png.

Depth Mapping

To excecute the stereo vision depth mapping:

1. cd exp2_depthMapping
2. Run dot_detection_CONV.py and dot_detection_cv.py for guassian dot detection using a convolution method and openCV.
3. Run dot_calibration.py with calibration images to generate polynomial fit.
4. Run depth_mapping_cv.py and pass the local location of the desired images.
5. Run depth_mapping_fixed.py for single depth map. Note that this method takes approximately 10 minutes. Errors appear when using other images as the window size and grid need to be adjusted manually.

Moth Eye image analysis

1. cd extensionMoth
2. Run circles.py, toggle DIsplay to see statistics plots.