table of Contents
0. mounting plate recognition
https://blog.csdn.net/lilai619/article/details/79695109
1.git clone-date documentation
cd /opt/deploy
git clone https://github.com/sergiomsilva/alpr-unconstrained.git
2. Download the pre-training model
cd alpr-unconstrained/
bash get-networks.sh
3. Delete the project comes Darknet replaced by official comes darknet
rm -rf darknet
git clone https://github.com/pjreddie/darknet
4. Change the official darknet support gpu and make due to my cudn drive installed in the default location so I just need to change three
cd darknet/
vim Makefile
将第1、2行的 已支持GPU
GPU=0
CUDNN=0
修改成：
GPU=1
CUDNN=1
将24行的 支持 cudnn
NVCC=nvcc
修改成：
NVCC=/usr/local/cuda/bin/nvcc
:wq
5. Compile
make all -j6 根据自己核数调整
6. Re-enter the main directory
cd /opt/alpr-unconstrained
cp -R data/* darknet/data/
7. Change file names
vim data/ocr/ocr-net.names
0
1
2
3
4
5
6
7
8
9
A
B
C
D
E
F
G
H
I
J
K
L
M
N
P
Q
R
S
T
U
V
W
X
Y
Z
京
沪
津
渝
冀
晋
蒙
辽
吉
黑
苏
浙
皖
闽
赣
鲁
豫
鄂
湘
粤
桂
琼
川
贵
云
藏
陕
甘
青
宁
新
Position changes corresponding class training files are located
vim data/ocr/ocr-net.data
classes=66
names=data/ocr/ocr-net.names
train=data/ocr/train.txt
backup=data/ocr/output
Create output directory
mkdir -p data/ocr/output
CFG layers and modify training parameters of the network layer
cp /opt/deploy/darknet/cfg/yolov3.cfg data/ocr/ocr-net.cfg
vim data/ocr/ocr-net.cfg
据自己GPU 和内存来指定 cfg 部分
训练的时候将第3、4行的 已支持GPU
batch=64
subdivisions=4
[net]
# Testing
# batch=1
# subdivisions=1
# Training
batch=64
subdivisions=8
......
[convolutional]
size=1
stride=1
pad=1
filters=33###75
activation=linear
[yolo]
mask = 6,7,8
anchors = 10,13, 16,30, 33,23, 30,61, 62,45, 59,119, 116,90, 156,198, 373,326
classes=6###20
num=9
jitter=.3
ignore_thresh = .5
truth_thresh = 1
random=0###1
......
[convolutional]
size=1
stride=1
pad=1
filters=33###75
activation=linear
[yolo]
mask = 3,4,5
anchors = 10,13, 16,30, 33,23, 30,61, 62,45, 59,119, 116,90, 156,198, 373,326
classes=6###20
num=9
jitter=.3
ignore_thresh = .5
truth_thresh = 1
random=0###1
......
[convolutional]
size=1
stride=1
pad=1
filters=33###75
activation=linear
[yolo]
mask = 0,1,2
anchors = 10,13, 16,30, 33,23, 30,61, 62,45, 59,119, 116,90, 156,198, 373,326
classes=6###20
num=9
jitter=.3
ignore_thresh = .5
truth_thresh = 1
random=0###1
————————————
filters数目是怎么计算的：3x(classes数目+5)，和聚类数目分布有关，论文中有说明；
比如说我有66类 那么 3* （66+5）=213
1. A method of correcting plate
https://github.com/zeusees/HyperLPR
part of the source code (I added a point comment)
def findContoursAndDrawBoundingBox(image_rgb):
line_upper = [];
line_lower = [];
line_experiment = []
grouped_rects = []
gray_image = cv2.cvtColor(image_rgb,cv2.COLOR_BGR2GRAY)
# for k in np.linspace(-1.5, -0.2,10):
for k in np.linspace(-50, 0, 15):
# thresh_niblack = threshold_niblack(gray_image, window_size=21, k=k)
# binary_niblack = gray_image > thresh_niblack
# binary_niblack = binary_niblack.astype(np.uint8) * 255
# 当一幅图像上的不同部分具有不同亮度时,我们需要采用自适应阈值.此时的阈值是根据图像上的每一个小区域计算与其
# 对应的阈值.因此,在同一幅图像上的不同区域采用的是不同的阈值,从而使我们能在亮度不同的情况下得到更好的结果.
"""
Args:
- src, 原图像,应该是灰度图
- x, 指当像素值高于(有时是低于)阈值时应该被赋予新的像素值, 255是白色
- adaptive_method, CV_ADAPTIVE_THRESH_MEAN_C 或 CV_ADAPTIVE_THRESH_GAUSSIAN_C
- threshold_type: 指取阈值类型
. CV_THRESH_BINARY, 二进制阈值化
. CV_THRESH_BINARY_INV, 反二进制阈值化
- block_size: 用来计算阈值的像素邻域大小(块大小):3,5,7,...
- param1: 指与方法有关的参数.对方法CV_ADAPTIVE_THRESH_MEAN_C和CV_ADAPTIVE_THRESH_GAUSSIAN_C,它是一个从均值或加权均值提取的常数,尽管它可以是负数。
. 对方法 CV_ADAPTIVE_THRESH_MEAN_C，先求出块中的均值，再减掉param1。
. 对方法 CV_ADAPTIVE_THRESH_GAUSSIAN_C ，先求出块中的加权和(gaussian)， 再减掉param1。
"""
binary_niblack = cv2.adaptiveThreshold(gray_image,255,cv2.ADAPTIVE_THRESH_MEAN_C,cv2.THRESH_BINARY,17,k) #邻域大小17是不是太大了??
#cv2.imshow("image1",binary_niblack)
#cv2.waitKey(0)
#imagex, contours, hierarchy = cv2.findContours(binary_niblack.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE)
contours, hierarchy = cv2.findContours(binary_niblack.copy(),cv2.RETR_EXTERNAL,cv2.CHAIN_APPROX_SIMPLE) # modified by bigz
for contour in contours:
#用一个最小的矩形,把找到的形状包起来
bdbox = cv2.boundingRect(contour)
if (bdbox[3]/float(bdbox[2])>0.7 and bdbox[3]*bdbox[2]>100 and bdbox[3]*bdbox[2]<1200) or (bdbox[3]/float(bdbox[2])>3 and bdbox[3]*bdbox[2]<100):
# cv2.rectangle(rgb,(bdbox[0],bdbox[1]),(bdbox[0]+bdbox[2],bdbox[1]+bdbox[3]),(255,0,0),1)
line_upper.append([bdbox[0],bdbox[1]])
line_lower.append([bdbox[0]+bdbox[2],bdbox[1]+bdbox[3]])
line_experiment.append([bdbox[0],bdbox[1]])
line_experiment.append([bdbox[0]+bdbox[2],bdbox[1]+bdbox[3]])
# grouped_rects.append(bdbox)
"""
想为图像周围建一个边使用訪函数,这经常在卷积运算或0填充时被用到
Args:
- src: 输入图像
- top,bottom,left,right 对应边界的像素数目
- borderType: 要添加哪种类型的边界
- BORDER_CONSTANT #边缘填充用固定像素值,比如填充黑边,就用0,白边255
- BORDER_REPLICATE #用原始图像相应的边缘的像素去做填充,看起来有一种把图像边缘"拉糊了"的效果
"""
rgb = cv2.copyMakeBorder(image_rgb,30,30,0,0,cv2.BORDER_REPLICATE)
leftyA, rightyA = fitLine_ransac(np.array(line_lower),3)
rows,cols = rgb.shape[:2]
# rgb = cv2.line(rgb, (cols - 1, rightyA), (0, leftyA), (0, 0, 255), 1,cv2.LINE_AA)
leftyB, rightyB = fitLine_ransac(np.array(line_upper),-3)
rows,cols = rgb.shape[:2]
# rgb = cv2.line(rgb, (cols - 1, rightyB), (0, leftyB), (0,255, 0), 1,cv2.LINE_AA)
pts_map1 = np.float32([[cols - 1, rightyA], [0, leftyA],[cols - 1, rightyB], [0, leftyB]])
pts_map2 = np.float32([[136,36],[0,36],[136,0],[0,0]])
mat = cv2.getPerspectiveTransform(pts_map1,pts_map2)
image = cv2.warpPerspective(rgb,mat,(136,36),flags=cv2.INTER_CUBIC)
#校正角度
#cv2.imshow("校正前",image)
#cv2.waitKey(0)
image,M = deskew.fastDeskew(image)
#cv2.imshow("校正后",image)
#cv2.waitKey(0)
return image
