Skip to content

olpotkin/Vehicle-Detection

Repository files navigation

Vehicle Detection and Tracking Project

The goals / steps of this project are the following:

  • Perform a Histogram of Oriented Gradients (HOG) feature extraction on a labeled training set of images and train a classifier Linear SVM classifier.
  • Apply a color transform and append binned color features, as well as histograms of color, to the HOG feature vector.
  • For those first two steps requred normalization of features and randomization of a selection for training and testing.
  • Implement a sliding-window technique and use trained classifier to search for vehicles in images.
  • Run pipeline on a video stream and create a heat map of recurring detections frame by frame to reject outliers and follow detected vehicles.
  • Estimate a bounding box for vehicles detected.

Rubric Points

All code for this project stored in Vehicle-Detection.ipynb IPython Notebook.

Code consists of 2 main Pipelines:

  • Lane Lines Detection (cells 1 - 15)
  • Vehicle Detection (cells 16 - 22)

0. Dataset

Dataset includes:

  • 8792 images with cars
  • 8968 images without cars

Histogram of Oriented Gradients (HOG)

1. Extract HOG features from the training images.

The code for this step is contained in the code cell #16 of the IPython notebook.

Reading in all the vehicle and non-vehicle images. Here is an example of one of each of the vehicle and non-vehicle classes:

alt text

Explore different color spaces and different skimage.hog() parameters (orientations, pixels_per_cell, and cells_per_block). I grabbed random images from each of the two classes and displayed them to get a feel for what the skimage.hog() output looks like.

Here is an example using the YCrCb color space and HOG parameters of orientations=9, pixels_per_cell=(8, 8) and cells_per_block=(2, 2):

alt text

2. Final choice of HOG parameters

Various combinations of parameters were tried and settled with a stable set:

color_space    = 'YCrCb'    # Can be RGB, HSV, LUV, HLS, YUV, YCrCb
orient         = 9          # HOG orientations
pix_per_cell   = 8          # HOG pixels per cell
cell_per_block = 2          # HOG cells per block
hog_channel    = 'ALL'      # Can be 0, 1, 2, or "ALL"
spatial_size   = (32, 32)   # Spatial binning dimensions
hist_bins      = 32         # Number of histogram bins
spatial_feat   = True       # Spatial features on or off
hist_feat      = True       # Histogram features on or off
hog_feat       = True       # HOG features on or off

Tried different color spaces, but YCrCb shown best results.

Increasing the orientation enhanced the accuarcy of the classifier, but increased computational time.

3. Train a classifier using selected HOG features and color features.

The code for this step is contained in the code cell #18 of the IPython notebook.

The extracted features where fed to LinearSVC model of sklearn with default settings. The trained model had accuracy of 99.35% on test dataset.

The trained model and parameters used for training were saved to pickle file to be further used by vehicle detection pipeline.

Sliding Window Search

1. Sliding window search.

For higher coverage of potential detections the multi-scale window approach was used. It's prevents calculation of feature vectors for the complete image and thus helps in speeding up the process.

Scale 1 Scale 2 Scale 3
ystart = 380 ystart = 400 ystart = 500
ystop = 480 ystop = 600 ystop = 700
scale = 1 scale = 1.5 scale = 2.5

The figure below shows the multiple scales under consideration overlapped on image.

alt text

2. Examples of test images to demonstrate how pipeline is working.

Ultimately I searched on 3 scales using YCrCb 3-channel HOG features plus spatially binned color and histograms of color in the feature vector, which provided a nice result.

I saved the positions of true detections in each frame of the video. From the true detections I created a heatmap and then thresholded that map to identify vehicle positions:

alt text

Final step is to join 2 pipelines: Lane Line Detection And Vehicle Detection:

alt text

Video Implementation

Here's a link to my video result on YouTube


Discussion

  • Neural Network could show a higher precision.

  • Pipeline may have problems in difficult lighting and illumination conditions.

  • The multi-window search may be optimized further for better speed and accuracy.