-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathrealtime.py
171 lines (116 loc) · 5.25 KB
/
realtime.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
import torch
from screen_shot import Capturer,Timer
import time
import cv2
from win32con import HWND_TOPMOST, SWP_NOMOVE, SWP_NOSIZE # win32core ==>> win32con
from win32gui import FindWindow, SetWindowPos
import numpy as np
import pydirectinput
import keyboard
import pyautogui
from logitech import Logitech
import pandas as pd
pyautogui.FAILSAFE=False
model=torch.hub.load('yolov5-master','custom','yolov5-master/runs/train/exp21/weights/best.pt',source="local")# use "custom" else it will download .pt from github.
#参数设置：
model.classes=[0] #coco.yaml
model.conf=0.35
print(model.conf)
region=region = (0, 0,2560, 1440)
title="Visualized Image"
cv2.namedWindow("Visualized Image", cv2.WINDOW_NORMAL)
cv2.resizeWindow("Visualized Image", 320, 180)
cv2.setWindowProperty("Visualized Image", cv2.WND_PROP_TOPMOST, 1)
#鼠标移动
origion_pos=[1280,720]
def mouse_track(lr,rb):
#x, y = pyautogui.position() # 返回鼠标的坐标
x=origion_pos[0]
y=origion_pos[1]
dx = lr - x
dy = rb - y
Logitech.mouse.move(dx//4,dy//4)
Logitech.mouse.move(dx // 4, dy // 4)
Logitech.mouse.move(dx // 4, dy // 4)
Logitech.mouse.move(dx // 4, dy // 4)
print("鼠标移动量:",dx," ",dy)
return dx,dy
import ctypes
import os
import pynput
import winsound
#相对位移尝试用pydirectinput.move() move_to不对劲！
#储存移动量与位置 数据
data_dx=[]
data_dy=[]
data_targert_x=[]
data_targert_y=[]
#================
while True:
t1 = time.perf_counter_ns()
img = Capturer.backup(region)
t2 = time.perf_counter_ns()
#BGR==>RGB
img = model(img)
#==========
#print(type(img))
#img = img.ims[0]
#print(dir(img))
# 获取处理后的图像和检测结果
#processed_image = img.ims[0]
# 获取处理后的图像和检测结果
#print(img)
processed_image = img.ims[0]
detections = img.pred[0]
# 复制图像以避免修改原始图像
result_image = np.copy(processed_image)
# 遍历检测结果并绘制边界框
# 遍历检测结果并绘制边界框 边界框始终画不对！！！！！！！！！
dxdyes=[] #store the distance from centre to target box.
box_position=[] #store the position of boxes
for det in detections:
left, top, right, bottom, confidence, class_id = det[0:6].cpu().numpy()
left, top, right, bottom = int(left), int(top), int(right), int(bottom)
color = (0, 255, 0) # 边界框颜色，这里使用绿色
label = f'Class: {int(class_id)}, Conf: {confidence:.2f}' # 类别和置信度信息
#print("坐标:",left,top,right,bottom)
cv2.rectangle(result_image, (left, top), (right, bottom), color, 2)
cv2.putText(result_image, label, (left, top - 10), cv2.FONT_HERSHEY_SIMPLEX, 0.5, color, 2)
lr = (left + right) // 2
rb = (top + bottom) // 2
box_position.append([lr,rb])
dxdyes.append((lr-origion_pos[0])**2 + (rb-origion_pos[1])**2)#距离最小准测 #add the linear distance. so each box has this number in dxdyes
#dxdyes.append((right-left)*(bottom-top))#面积最大准测,这里取负数是为了跟下面的min切合,有点问题，效果不太好。
print("conf=",confidence)
#and we just find the minium distance by the box(I mean the box that is closest to our centre)
if dxdyes!=[]:
print("距离/面积：",dxdyes)
min_index=dxdyes.index(min(dxdyes))
print("目标起始位置:",box_position[min_index][0],box_position[min_index][1])
if keyboard.is_pressed('alt'):
dx,dy=mouse_track(box_position[min_index][0],box_position[min_index][1])
#data_dx.append(dx)
# data_dy.append(dy)
#data_targert_x.append(box_position[min_index][0])
#data_targert_y.append(box_position[min_index][1])
# 显示可视化结果
cv2.imshow("Visualized Image", result_image)
k = cv2.waitKey(1) # 0:不自动销毁也不会更新, 1:1ms延迟销毁
if k % 256 == 27:
cv2.destroyAllWindows()
# 保存数据
# 创建一个字典，其中键是列名，值是数据
# data = {'dx': data_dx, 'dy': data_dy, 'target_x': data_targert_x, 'target_y': data_targert_y}
# 使用pandas创建一个DataFrame
#df = pd.DataFrame(data)
# 将DataFrame保存为csv文件
#df.to_csv('data_mouse.csv', index=False)
exit('ESC ...')
time.sleep(0.01)
#仍然存在的问题
#1.训练程度貌似不够？ 2.鼠标灵敏度与 pydirectinput.move的关系 3.瞄不准？
#4.鼠标 触底现象！ 鼠标达到边框后move失效！ 我们应该即使更新鼠标的位置！！！！
#5.灵敏度为8时定位不准 猜测可能原因1.识别不准（不太可能，因为灵敏度为1时是准的） 2.鼠标移动具有加速度 3.与灵敏度有关需要调整logitech.move的移动量？
#打算收集的数据 1.鼠标相对位移 此即dx，dy 2.目标初始位置/移动后位置 3.鼠标灵敏度
#MOUSE.MOVE 是以一个力来把鼠标抛出去，所以灵敏度不同会导致其不准，比如我设置的8倍数，就会抛过了；我打算收集一下鼠标移动量 目标位置 目标是测出准星的移动量与鼠标移动量的关系？
#有加速度，不是单纯的8倍关系 唉，想放弃了