Initial commit of winning code

detect_pedestrians.py

@@ -0,0 +1,89 @@
+#!/usr/bin/python
+# -*- coding: utf-8 -*-
+
+__author__ = 'nikolajus krauklis <[email protected]>'
+
+import os
+import sys
+from optparse import OptionParser
+import time
+
+import cv2
+
+from pedestrian_lib import *
+
+
+usage = "usage: %prog [options] pictures.txt"
+parser = OptionParser(usage=usage, version="%prog 1.0.1")
+
+parser.add_option("-d", "--debug", action="store_true",
+                  help="show debugging window and do not create blob files", default=False)
+parser.add_option("-v", "--verbose", action="store_true", default=False,
+                  help="show more information in stdout")
+parser.add_option("-c", "--classifier", default="haar_classifier.xml",
+                  help="classifier file name. default: haar_classifier.xml")
+parser.add_option("-l", "--debug-level", default=1, dest="debug_level")
+
+(options, args) = parser.parse_args()
+
+DEBUG = options.debug
+DEBUG_LEVEL = options.debug_level
+VERBOSE = options.verbose
+
+
+def process(images, cascade):
+    for image in images:
+        started_process = time.time()
+
+        if os.path.exists(image):
+            cv_image = cv2.imread(image)
+        else:
+            if VERBOSE:
+                print "image file '{0}' does not exists. skipping...".format(image)
+            continue
+
+        haar_objects = detect_objects(cv_image, cascade)
+        color_blobs = detect_blobs(cv_image, DEBUG, DEBUG_LEVEL)
+        final_blobs = []
+
+        for haar_object in haar_objects:
+            found = False
+            for color_blob in color_blobs:
+                if is_overlapping(haar_object, color_blob):
+                    found = True
+                    break
+            if found:
+                final_blobs.append(haar_object)
+
+        elapsed_time = time.time() - started_process
+
+        if VERBOSE:
+            print "{0} pedestrian traffic signs found in image '{1}' in {2}s".format(len(final_blobs), image,
+                                                                                     elapsed_time)
+            print final_blobs
+
+        if not DEBUG:
+            save_found_blobs(image, final_blobs)
+        else:
+            show_blobs(cv_image, final_blobs)
+
+
+if __name__ == '__main__':
+    if len(args) == 0 or not os.path.exists(args[0]):
+        print_usage()
+        parser.print_help()
+        sys.exit(-1)
+
+    file_name = args[0]
+    classifier_filename = options.classifier
+
+    images = read_image_list(file_name)
+
+    if not os.path.exists(classifier_filename):
+        print "Classifier file does not exist"
+        parser.print_help()
+        sys.exit(-1)
+
+    cascade = cv2.CascadeClassifier(classifier_filename)
+
+    process(images, cascade)