1 Star 1 Fork 90

mgbin / hand-keras-yolo3-recognize

加入 Gitee
与超过 1200万 开发者一起发现、参与优秀开源项目,私有仓库也完全免费 :)
免费加入
克隆/下载
yolo.py 15.11 KB
一键复制 编辑 原始数据 按行查看 历史
cungudafa 提交于 2020-08-08 09:50 . sign
# -*- coding: utf-8 -*-
"""
Class definition of YOLO_v3 style detection model on image and video
"""
import colorsys
from timeit import default_timer as timer
import os
import numpy as np
import tensorflow as tf
from tensorflow.keras.models import load_model
from tensorflow.keras.layers import Input
from tensorflow.keras.utils import multi_gpu_model
from PIL import Image, ImageFont, ImageDraw
from yolo3.model import yolo_eval, yolo_body, tiny_yolo_body
from yolo3.utils import letterbox_image
from pose.coco import general_coco_model
class YOLO(object):
_defaults = {
"model_path": 'D:/myworkspace/JupyterNotebook/hand-keras-yolo3-recognize/model/yolov3/last1.h5', # 模型
"anchors_path": 'D:/myworkspace/JupyterNotebook/hand-keras-yolo3-recognize/model/yolov3/coco_anchors.txt', # 先验框
"classes_path": 'D:/myworkspace/JupyterNotebook/hand-keras-yolo3-recognize/model/yolov3/voc_classes.txt', # 种类
"score": 0.3, # 框置信度阈值,小于阈值的框被删除,需要的框较多,则调低阈值,需要的框较少,则调高阈值
"iou": 0.45, # 同类别框的IoU阈值,大于阈值的重叠框被删除,重叠物体较多,则调高阈值,重叠物体较少,则调低阈值
"model_image_size": (416, 416),
"gpu_num": 1,
}
@classmethod
def get_defaults(cls, n):
if n in cls._defaults:
return cls._defaults[n]
else:
return "Unrecognized attribute name '" + n + "'"
def __init__(self, **kwargs):
self.__dict__.update(self._defaults) # set up default values
self.__dict__.update(kwargs) # and update with user overrides
self.class_names = self._get_class()
self.anchors = self._get_anchors()
self.load_yolo_model()
def _get_class(self):
classes_path = os.path.expanduser(self.classes_path)
with open(classes_path) as f:
class_names = f.readlines()
class_names = [c.strip() for c in class_names]
return class_names
def _get_anchors(self):
anchors_path = os.path.expanduser(self.anchors_path)
with open(anchors_path) as f:
anchors = f.readline()
anchors = [float(x) for x in anchors.split(',')]
return np.array(anchors).reshape(-1, 2)
def load_yolo_model(self):
model_path = os.path.expanduser(self.model_path)
assert model_path.endswith(
'.h5'), 'Keras model or weights must be a .h5 file.'
# Load model, or construct model and load weights.
num_anchors = len(self.anchors)
num_classes = len(self.class_names)
is_tiny_version = num_anchors == 6 # default setting
try:
self.yolo_model = load_model(model_path, compile=False)
except:
self.yolo_model = tiny_yolo_body(Input(shape=(None, None, 3)), num_anchors // 2, num_classes) \
if is_tiny_version else yolo_body(Input(shape=(None, None, 3)), num_anchors // 3, num_classes)
# make sure model, anchors and classes match
self.yolo_model.load_weights(self.model_path)
else:
assert self.yolo_model.layers[-1].output_shape[-1] == \
num_anchors / len(self.yolo_model.output) * (num_classes + 5), \
'Mismatch between model and given anchor and class sizes'
#print('{} model, anchors, and classes loaded.'.format(model_path))
# Generate colors for drawing bounding boxes.
hsv_tuples = [(x / len(self.class_names), 1., 1.)
for x in range(len(self.class_names))]
self.colors = list(map(lambda x: colorsys.hsv_to_rgb(*x), hsv_tuples))
self.colors = list(
map(lambda x: (int(x[0] * 255), int(x[1] * 255), int(x[2] * 255)),
self.colors))
np.random.seed(10101) # Fixed seed for consistent colors across runs.
# Shuffle colors to decorrelate adjacent classes.
np.random.shuffle(self.colors)
np.random.seed(None) # Reset seed to default.
def compute_output(self, image_data, image_shape):
# Generate output tensor targets for filtered bounding boxes.
# self.input_image_shape = K.placeholder(shape=(2,))
self.input_image_shape = tf.constant(image_shape)
if self.gpu_num >= 2:
self.yolo_model = multi_gpu_model(
self.yolo_model, gpus=self.gpu_num)
boxes, scores, classes = yolo_eval(self.yolo_model(image_data), self.anchors,
len(self.class_names), self.input_image_shape,
score_threshold=self.score, iou_threshold=self.iou)
return boxes, scores, classes
def getyoloPoints(self, image):
"""yolo手部关键点检测
:param 图像
:return 信息[手的个数,结果,左上角坐标,右下角坐标,...]
"""
start = timer()
if self.model_image_size != (None, None):
assert self.model_image_size[0] % 32 == 0, 'Multiples of 32 required'
assert self.model_image_size[1] % 32 == 0, 'Multiples of 32 required'
boxed_image = letterbox_image(
image, tuple(reversed(self.model_image_size))) # 原图转换成数组格式
else:
new_image_size = (image.width - (image.width % 32),
image.height - (image.height % 32))
boxed_image = letterbox_image(image, new_image_size)
image_data = np.array(boxed_image, dtype='float32')
image_data /= 255.
image_data = np.expand_dims(image_data, 0) # Add batch dimension.
out_boxes, out_scores, out_classes = self.compute_output(
image_data, [image.size[1], image.size[0]]) # yolo检测结果
if len(out_boxes) > 0:
print('Found {} boxes for {}'.format(len(out_boxes), 'img'))
Info = [] # 存放信息的列表
Info.append(len(out_boxes))
for i, c in reversed(list(enumerate(out_classes))):
box = out_boxes[i]
score = out_scores[i]
top, left, bottom, right = box
top = max(0, np.floor(top + 0.5).astype('int32'))
left = max(0, np.floor(left + 0.5).astype('int32'))
bottom = min(image.size[1], np.floor(bottom + 0.5).astype('int32'))
right = min(image.size[0], np.floor(right + 0.5).astype('int32'))
# print(label, (left, top), (right, bottom))#打印检测结果、左上角坐标和右下角坐标
Info.append(format(score))
Info.append((left, top))
Info.append((right, bottom))
end = timer()
if len(out_boxes) > 0:
print('[INFO]yolo_Model predicts time: {}'.format(end - start))
return Info
def vis_hand_pose(self, img_cv2, dotimg, black_np,bone_points, yololabel):
from cv2 import cv2
"""显示yolo寻找到手势位置的图像
:param 图像,yolo信息包含检测关键点坐标
:return 手势框
"""
#img_old = cv2.imread(imgfile)
points = []
if yololabel[0] == 1: # 一只手
points.append(yololabel[2])
points.append(yololabel[3])
# cvRectangle函数参数: 图片, 左上角, 右下角, 颜色, 线条粗细, 线条类型,点类型
cv2.rectangle(img_cv2, yololabel[2],
yololabel[3], (255, 0, 0), 1, 4, 0 ) # 画yolo矩形
cv2.rectangle(black_np, yololabel[2],
yololabel[3], (255, 0, 0), 1, 4, 0 )
cv2.rectangle(dotimg, yololabel[2],
yololabel[3], (255, 0, 0), 1, 4, 0 )
if yololabel[0] > 1: # 2只手
points.append(yololabel[2])
points.append(yololabel[3])
cv2.rectangle(img_cv2, yololabel[2], yololabel[3], (255, 0, 0), 1, 4, 0 )
cv2.rectangle(black_np, yololabel[2],
yololabel[3], (255, 0, 0), 1, 4, 0 )
cv2.rectangle(dotimg, yololabel[2],
yololabel[3], (255, 0, 0), 1, 4, 0 )
points.append(yololabel[5])
points.append(yololabel[6])
cv2.rectangle(img_cv2, yololabel[5], yololabel[6], (255, 0, 0), 1, 4, 0 )
cv2.rectangle(black_np, yololabel[5],
yololabel[6], (255, 0, 0), 1, 4, 0 )
cv2.rectangle(dotimg, yololabel[5],
yololabel[6], (255, 0, 0), 1, 4, 0 )
for idx,yolopoint in enumerate(points):
if yolopoint:
# 线
cv2.line(black_np, yolopoint,
bone_points[1], (0, 255, 0), 2) # 手与骨骼点1的连线
cv2.line(dotimg, yolopoint,
bone_points[1], (0, 255, 0), 2)
cv2.line(img_cv2, yolopoint,
bone_points[1], (0, 255, 0), 2)
# 点
cv2.circle(
black_np, yolopoint, 3, (255, 0, 0), thickness=-1, lineType=cv2.FILLED) # yolo手点
cv2.circle(
img_cv2, yolopoint, 3, (255, 0, 0), thickness=-1, lineType=cv2.FILLED)
cv2.circle(
dotimg, yolopoint, 3, (255, 0, 0), thickness=-1, lineType=cv2.FILLED)
# 骨骼点1
cv2.circle(
img_cv2, bone_points[1], 3, (255, 0, 0), thickness=-1, lineType=cv2.FILLED)
cv2.circle(
dotimg, bone_points[1], 3, (255, 0, 0), thickness=-1, lineType=cv2.FILLED)
cv2.circle(
black_np, bone_points[1], 3, (255, 0, 0), thickness=-1, lineType=cv2.FILLED)
return img_cv2, dotimg, black_np # 连线图,点图,黑连线图
def detect_image(self, image, lineimage):
start = timer()
if self.model_image_size != (None, None):
assert self.model_image_size[0] % 32 == 0, 'Multiples of 32 required'
assert self.model_image_size[1] % 32 == 0, 'Multiples of 32 required'
boxed_image = letterbox_image(
image, tuple(reversed(self.model_image_size))) # 原图转换成数组格式
else:
new_image_size = (image.width - (image.width % 32),
image.height - (image.height % 32))
boxed_image = letterbox_image(image, new_image_size)
image_data = np.array(boxed_image, dtype='float32')
image_data /= 255.
image_data = np.expand_dims(image_data, 0) # Add batch dimension.
out_boxes, out_scores, out_classes = self.compute_output(
image_data, [image.size[1], image.size[0]]) # yolo检测结果
if len(out_boxes) > 0:
print('Found {} boxes for {}'.format(len(out_boxes), 'img'))
font = ImageFont.truetype(font='docs/font/FiraMono-Medium.otf',
size=np.floor(3e-2 * image.size[1] + 0.5).astype('int32'))
thickness = (image.size[0] + image.size[1]) // 300
Info = [] # 存放信息的列表
hand_ROI_PIL = [] # 存放手部图片
Info.append(len(out_boxes))
for i, c in reversed(list(enumerate(out_classes))):
predicted_class = self.class_names[c]
box = out_boxes[i]
score = out_scores[i]
label = '{} {:.2f}'.format(predicted_class, score) # 打印标签类型、相似度
draw = ImageDraw.Draw(lineimage) # 画在点线图上
label_size = draw.textsize(label, font)
top, left, bottom, right = box
top = max(0, np.floor(top + 0.5).astype('int32'))
left = max(0, np.floor(left + 0.5).astype('int32'))
bottom = min(image.size[1], np.floor(bottom + 0.5).astype('int32'))
right = min(image.size[0], np.floor(right + 0.5).astype('int32'))
# print(label, (left, top), (right, bottom))#打印检测结果、左上角坐标和右下角坐标
Info.append(format(score))
Info.append((left, top))
Info.append((right, bottom))
# 从左上角开始 剪切图片
img2 = image.crop((left, top, right, bottom)) # 原图中手的图片,处理
# img2.save("docs/"+label+"_"+str(i)+".jpg")
hand_ROI_PIL.append(img2)
# 画图
if top - label_size[1] >= 0:
text_origin = np.array([left, top - label_size[1]])
else:
text_origin = np.array([left, top + 1])
# My kingdom for a good redistributable image drawing library.
for i in range(thickness):
draw.rectangle(
[left + i, top + i, right - i, bottom - i],
outline=self.colors[c]) # 在点线图中框出手,显示
draw.rectangle(
[tuple(text_origin), tuple(text_origin + label_size)],
fill=self.colors[c]) # 写hand识别率的框
draw.text(text_origin, label, fill=(
0, 0, 0), font=font) # 写上标签和识别率
del draw
end = timer()
if len(out_boxes) > 0:
print('[INFO]yolo_Model predicts time: {}'.format(end - start))
return lineimage, Info, hand_ROI_PIL
def detect_video(yolo, video_path, output_path=""):
from cv2 import cv2
vid = cv2.VideoCapture(video_path)
if not vid.isOpened():
raise IOError("Couldn't open webcam or video")
video_FourCC = int(vid.get(cv2.CAP_PROP_FOURCC)) # 获取原始视频的信息
video_fps = vid.get(cv2.CAP_PROP_FPS)
video_size = (int(vid.get(cv2.CAP_PROP_FRAME_WIDTH)),
int(vid.get(cv2.CAP_PROP_FRAME_HEIGHT)))
isOutput = True if output_path != "" else False # 如果设置了视频保存路径,则保存视频
if isOutput:
print("!!! TYPE:", type(output_path), type(
video_FourCC), type(video_fps), type(video_size))
out = cv2.VideoWriter(output_path, video_FourCC,
video_fps, video_size) # 根据原视频设置 保存视频的路径、大小、帧数
accum_time = 0
curr_fps = 0
fps = "FPS: ??"
prev_time = timer()
while True:
return_value, frame = vid.read()
image = Image.fromarray(frame)
lineimage, Info, hand_ROI_PIL = yolo.detect_image(image,image) # 检测
result = np.asarray(Image.fromarray(lineimage)) # 画图到全部图上
curr_time = timer()
exec_time = curr_time - prev_time
prev_time = curr_time
accum_time = accum_time + exec_time
curr_fps = curr_fps + 1
if accum_time > 1:
accum_time = accum_time - 1
fps = "FPS: " + str(curr_fps)
curr_fps = 0
cv2.putText(result, text=fps, org=(3, 15), fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.50, color=(255, 0, 0), thickness=2)
cv2.putText(result, "q-'quit'", org=(3, 45), fontFace=cv2.FONT_HERSHEY_SIMPLEX,
fontScale=0.50, color=(0, 255, 0), thickness=2) # 标注字体
cv2.namedWindow("result", cv2.WINDOW_NORMAL)
cv2.imshow("result", result)
if isOutput:
out.write(result)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
Python
1
https://gitee.com/mgb/hand-keras-yolo3-recognize.git
git@gitee.com:mgb/hand-keras-yolo3-recognize.git
mgb
hand-keras-yolo3-recognize
hand-keras-yolo3-recognize
master

搜索帮助