From fd0b621615288ac78741ae9edfe7abe736698b58 Mon Sep 17 00:00:00 2001
From: Edmond Yoo <hj3yoo@uwaterloo.ca>
Date: Fri, 14 Sep 2018 23:18:07 +0000
Subject: [PATCH] Glare spotting & correction
---
opencv_dnn.py | 140 +++++++++++++++++++++++++++++++++-------------
1 files changed, 99 insertions(+), 41 deletions(-)
diff --git a/opencv_dnn.py b/opencv_dnn.py
index 8746c84..9b4c1ad 100644
--- a/opencv_dnn.py
+++ b/opencv_dnn.py
@@ -2,6 +2,7 @@
import numpy as np
import os
import sys
+from operator import itemgetter
# Disclaimer: majority of the basic framework in this file is modified from the following tutorial:
@@ -17,7 +18,7 @@
# Remove the bounding boxes with low confidence using non-maxima suppression
-def postprocess(frame, outs, classes, thresh_conf, thresh_nms):
+def post_process(frame, outs, thresh_conf, thresh_nms):
frame_height = frame.shape[0]
frame_width = frame.shape[1]
@@ -42,17 +43,11 @@
confidences.append(float(confidence))
boxes.append([left, top, width, height])
- # Perform non maximum suppression to eliminate redundant overlapping boxes with
- # lower confidences.
- indices = cv2.dnn.NMSBoxes(boxes, confidences, thresh_conf, thresh_nms)
- for i in indices:
- i = i[0]
- box = boxes[i]
- left = box[0]
- top = box[1]
- width = box[2]
- height = box[3]
- draw_pred(frame, class_ids[i], classes, confidences[i], left, top, left + width, top + height)
+ # Perform non maximum suppression to eliminate redundant overlapping boxes with lower confidences.
+ indices = [ind[0] for ind in cv2.dnn.NMSBoxes(boxes, confidences, thresh_conf, thresh_nms)]
+
+ ret = [[class_ids[i], confidences[i], boxes[i]] for i in indices]
+ return ret
# Draw the predicted bounding box
@@ -73,7 +68,33 @@
cv2.putText(frame, label, (left, top), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255, 255, 255))
-def detect_frame(net, classes, img, thresh_conf=0.5, thresh_nms=0.4, in_dim=(416, 416), out_path=None):
+def remove_glare(img):
+ img_hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)
+ _, s, v = cv2.split(img_hsv)
+ non_sat = (s < 32) * 255 # Find all pixels that are not very saturated
+
+ # Slightly decrease the area of the non-satuared pixels by a erosion operation.
+ disk = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (3, 3))
+ non_sat = cv2.erode(non_sat.astype(np.uint8), disk)
+
+ # Set all brightness values, where the pixels are still saturated to 0.
+ v[non_sat == 0] = 0
+ # filter out very bright pixels.
+ glare = (v > 240) * 255
+
+ # Slightly increase the area for each pixel
+ glare = cv2.dilate(glare.astype(np.uint8), disk)
+ #glare = cv2.dilate(glare.astype(np.uint8), disk);
+
+ #corrected = cv2.inpaint(img, glare, 7, cv2.INPAINT_TELEA)
+ glare_reduced = np.ones((img.shape[0], img.shape[1], 3), dtype=np.uint8) * 200
+ glare = cv2.cvtColor(glare, cv2.COLOR_GRAY2BGR)
+ corrected = np.where(glare, glare_reduced, img)
+ return corrected
+
+
+def detect_frame(net, classes, img, thresh_conf=0.5, thresh_nms=0.4, in_dim=(416, 416), display=True, out_path=None):
+ img_copy = img.copy()
# Create a 4D blob from a frame.
blob = cv2.dnn.blobFromImage(img, 1 / 255, in_dim, [0, 0, 0], 1, crop=False)
@@ -84,7 +105,11 @@
outs = net.forward(get_outputs_names(net))
# Remove the bounding boxes with low confidence
- postprocess(img, outs, classes, thresh_conf, thresh_nms)
+ obj_list = post_process(img, outs, thresh_conf, thresh_nms)
+ for obj in obj_list:
+ class_id, confidence, box = obj
+ left, top, width, height = box
+ draw_pred(img, class_id, classes, confidence, left, top, left + width, top + height)
# Put efficiency information. The function getPerfProfile returns the
# overall time for inference(t) and the timings for each of the layers(in layersTimes)
@@ -94,13 +119,37 @@
if out_path is not None:
cv2.imwrite(out_path, img.astype(np.uint8))
+ if display:
+ no_glare = remove_glare(img_copy)
+ img_concat = np.concatenate((img, no_glare), axis=1)
+ cv2.imshow('result', img_concat)
+
+ '''
+ for i in range(len(obj_list)):
+ class_id, confidence, box = obj_list[i]
+ left, top, width, height = box
+ img_snip = img[max(0, top):min(img.shape[0], top + height), max(0, left):min(img.shape[1], left + width)]
+ #cv2.imshow('feature#%d' % i, img_snip)
+ img_hsv = cv2.cvtColor(img_snip, cv2.COLOR_BGR2HSV)
+ h, s, v = cv2.split(img_hsv)
+ #h = cv2.cvtColor(h, cv2.COLOR_GRAY2BGR)
+ s = cv2.cvtColor(s, cv2.COLOR_GRAY2BGR)
+ v = cv2.cvtColor(v, cv2.COLOR_GRAY2BGR)
+ img_concat = np.concatenate((img_snip, s, v), axis=1)
+ cv2.imshow('feature#%d - hsv' % i, img_concat)
+ '''
+ cv2.waitKey(0)
+ cv2.destroyAllWindows()
+
+ return obj_list
-def detect_video(net, classes, capture, thresh_conf=0.5, thresh_nms=0.4, in_dim=(416, 416), out_path=None):
+def detect_video(net, classes, capture, thresh_conf=0.5, thresh_nms=0.4, in_dim=(416, 416), display=True, out_path=None):
if out_path is not None:
vid_writer = cv2.VideoWriter(out_path, cv2.VideoWriter_fourcc('M', 'J', 'P', 'G'), 30,
(round(capture.get(cv2.CAP_PROP_FRAME_WIDTH)),
round(capture.get(cv2.CAP_PROP_FRAME_HEIGHT))))
+ max_num_obj = 0
while True:
ret, frame = capture.read()
if not ret:
@@ -108,28 +157,33 @@
print("End of video. Press any key to exit")
cv2.waitKey(0)
break
- '''
- # Create a 4D blob from a frame.
- blob = cv2.dnn.blobFromImage(frame, 1 / 255, in_dim, [0, 0, 0], 1, crop=False)
-
- # Sets the input to the network
- net.setInput(blob)
-
- # Runs the forward pass to get output of the output layers
- outs = net.forward(get_outputs_names(net))
-
- # Remove the bounding boxes with low confidence
- postprocess(frame, outs, classes, thresh_conf, thresh_nms)
-
- # Put efficiency information. The function getPerfProfile returns the
- # overall time for inference(t) and the timings for each of the layers(in layersTimes)
- t, _ = net.getPerfProfile()
- label = 'Inference time: %.2f ms' % (t * 1000.0 / cv2.getTickFrequency())
- cv2.putText(frame, label, (0, 15), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 0, 255))
- '''
- detect_frame(net, classes, frame,
- thresh_conf=thresh_conf, thresh_nms=thresh_nms, in_dim=in_dim, out_path=None)
- cv2.imshow('result', frame)
+ img = frame.copy()
+ obj_list = detect_frame(net, classes, frame, thresh_conf=thresh_conf, thresh_nms=thresh_nms, in_dim=in_dim,
+ display=False, out_path=None)
+ max_num_obj = max(max_num_obj, len(obj_list))
+ if display:
+ no_glare = remove_glare(img)
+ img_concat = np.concatenate((frame, no_glare), axis=1)
+ cv2.imshow('result', img_concat)
+ '''
+ for i in range(len(obj_list)):
+ class_id, confidence, box = obj_list[i]
+ left, top, width, height = box
+ img_snip = img[max(0, top):min(img.shape[0], top + height),
+ max(0, left):min(img.shape[1], left + width)]
+ # cv2.imshow('feature#%d' % i, img_snip)
+ img_hsv = cv2.cvtColor(img_snip, cv2.COLOR_BGR2HSV)
+ h, s, v = cv2.split(img_hsv)
+ # h = cv2.cvtColor(h, cv2.COLOR_GRAY2BGR)
+ s = cv2.cvtColor(s, cv2.COLOR_GRAY2BGR)
+ v = cv2.cvtColor(v, cv2.COLOR_GRAY2BGR)
+ img_concat = np.concatenate((img_snip, s, v), axis=1)
+ cv2.imshow('feature#%d - hsv' % i, img_concat)
+ for i in range(len(obj_list), max_num_obj):
+ cv2.imshow('feature#%d - hsv' % i, np.zeros((1, 1), dtype=np.uint8))
+ '''
+ #if len(obj_list) > 0:
+ #cv2.waitKey(0)
if out_path is not None:
vid_writer.write(frame.astype(np.uint8))
cv2.waitKey(1)
@@ -137,24 +191,27 @@
if out_path is not None:
vid_writer.release()
cv2.destroyAllWindows()
- pass
def main():
# Specify paths for all necessary files
- test_path = '../data/test1.mp4'
+ test_path = os.path.abspath('../data/test18.jpg')
weight_path = 'weights/second_general/tiny_yolo_final.weights'
cfg_path = 'cfg/tiny_yolo.cfg'
class_path = "data/obj.names"
out_dir = 'out'
if not os.path.isfile(test_path):
print('The test file %s doesn\'t exist!' % os.path.abspath(test_path))
+ return
if not os.path.isfile(weight_path):
print('The weight file %s doesn\'t exist!' % os.path.abspath(test_path))
+ return
if not os.path.isfile(cfg_path):
print('The config file %s doesn\'t exist!' % os.path.abspath(test_path))
+ return
if not os.path.isfile(class_path):
print('The class file %s doesn\'t exist!' % os.path.abspath(test_path))
+ return
# Setup
# Read class names from text file
@@ -162,8 +219,8 @@
classes = [line.strip() for line in f.readlines()]
# Load up the neural net using the config and weights
net = cv2.dnn.readNetFromDarknet(cfg_path, weight_path)
- #net.setPreferableBackend(cv2.dnn.DNN_BACKEND_OPENCV)
- #net.setPreferableTarget(cv2.dnn.DNN_TARGET_CPU)
+ net.setPreferableBackend(cv2.dnn.DNN_BACKEND_OPENCV)
+ net.setPreferableTarget(cv2.dnn.DNN_TARGET_CPU)
# Save the detection result if out_dir is provided
if out_dir is None or out_dir == '':
@@ -172,6 +229,7 @@
out_path = out_dir + '/' + os.path.split(test_path)[1]
# Check if test file is image or video
test_ext = test_path[test_path.find('.') + 1:]
+
if test_ext in ['jpg', 'jpeg', 'bmp', 'png', 'tiff']:
img = cv2.imread(test_path)
detect_frame(net, classes, img, out_path=out_path)
--
Gitblit v1.10.0