Week #22 6/25 - 7/2

Nick:

I spent the week working on the project report:

• imported chapters from proposal report

• updated tables and diagrams with the latest from the website

• created the structure of the report following the rubric

Ryan:

Continued troubleshooting code provided by Nick. Revised the code and added functionality for the COCO library to be used as the means of opening the door rather than the haar cascade image classifier.

Code:

import cv2

from gpiozero import Motor

from time import sleep

from tkinter import *

from tkinter import Tk, font

import datetime

import RPi.GPIO as GPIO

# GPIO setup

GPIO.setmode(GPIO.BCM)

GPIO.setup(17, GPIO.IN)

doormotor = Motor(24, 23)

#door roll up function

def rollup():

doormotor.forward()

sleep(7)

doormotor.stop()

#door roll down function with prox sensor added

def rolldown():

i = 0

while i < 7:

i = i+1

doormotor.backward()

sleep(1)

if not GPIO.input(17):

doormotor.forward()

sleep(2)

i = i-2

doormotor.stop()

#Full opening sequence function

def opensequence():

log_list.insert(END, datetime.datetime.now().strftime("%D - %H:%M:%S"))

rollup()

sleep(2)

rolldown()

#COCO object recognition code

classNames = []

classFile = "coco.names"

with open(classFile,"rt") as f:

classNames = f.read().rstrip("\n").split("\n")

configPath = "ssd_mobilenet_v3_large_coco_2020_01_14.pbtxt"

weightsPath = "frozen_inference_graph.pb"

net = cv2.dnn_DetectionModel(weightsPath,configPath)

net.setInputSize(320,320)

net.setInputScale(1.0/ 127.5)

net.setInputMean((127.5, 127.5, 127.5))

net.setInputSwapRB(True)

def getObjects(img, thres, nms, draw=True, objects=["dog"]):

classIds, confs, bbox = net.detect(img,confThreshold=thres,nmsThreshold=nms)

if len(objects) == 0: objects = classNames

objectInfo =[]

if len(classIds) != 0:

for classId, confidence,box in zip(classIds.flatten(),confs.flatten(),bbox):

className = classNames[classId - 1]

if className in objects:

objectInfo.append([box,className])

if draw:

cv2.rectangle(img,box,color=(0,255,0),thickness=2)

cv2.putText(img,classNames[classId-1].upper(),(box[0]+10,box[1]+30),

cv2.FONT_HERSHEY_COMPLEX,1,(0,255,0),2)

cv2.putText(img,str(round(confidence*100,2)),(box[0]+200,box[1]+30),

cv2.FONT_HERSHEY_COMPLEX,1,(0,255,0),2)

return img, objectInfo

# Camera setup

cam1 = cv2.VideoCapture(0)

cam1.set(cv2.CAP_PROP_FRAME_WIDTH, 800)

cam1.set(cv2.CAP_PROP_FRAME_HEIGHT, 600)

cam2 = cv2.VideoCapture(2)

cam2.set(cv2.CAP_PROP_FRAME_WIDTH, 800)

cam2.set(cv2.CAP_PROP_FRAME_HEIGHT, 600)

#function to determine if the current time is within the time window set by the sliders

def is_within_time_window(current_time):

start_time = datetime.time(tw_start_hour_slider.get(), tw_start_minute_slider.get())

end_time = datetime.time(tw_end_hour_slider.get(), tw_end_minute_slider.get())

return start_time <= current_time <= end_time

#scan function to determine if a dog is in the frame

def scan():

ret1, frame1 = cam1.read()

ret2, frame2 = cam2.read()

frame1_with_dogs, dogs_info1 = getObjects(frame1, 0.55, 0.2, objects=["dog"])

frame2_with_dogs, dogs_info2 = getObjects(frame2, 0.55, 0.2, objects=["dog"])

if dogs_info1 and is_within_time_window(datetime.datetime.now().time()):

opensequence()

if dogs_info2:

opensequence()

# GUI setup

main_window = Tk()

main_window.configure(background="SkyBlue4")

main_window.geometry("800x480")

main_window.attributes("-fullscreen", True)

main_window.title("Smart Doggy Door")

main_window.option_add("*font", "SegoeUI 24")

status = str("Locked")

tw_start = Label(main_window, bg='yellow', text='Time Window Start: ')

tw_start.grid(row=2, column=0)

tw_end = Label(main_window, bg='yellow', text='Time Window End: ')

tw_end.grid(row=3, column=0)

tw_start_hour_slider = Scale(main_window, label="HH", orient=HORIZONTAL, from_=0, to=23, background='SteelBlue3', font=('SegoeUI', 20))

tw_start_hour_slider.grid(row=2, column=1, sticky=W, padx=10, pady=10)

tw_start_minute_slider = Scale(main_window, label="MM", orient=HORIZONTAL, from_=0, to=59, background='SteelBlue3', font=('SegoeUI', 20))

tw_start_minute_slider.grid(row=2, column=2, sticky=W, padx=10, pady=10)

tw_end_hour_slider = Scale(main_window, label="HH", orient=HORIZONTAL, from_=0, to=23, background='SteelBlue3', font=('SegoeUI', 20))

tw_end_hour_slider.grid(row=3, column=1, sticky=W, padx=10, pady=10)

tw_end_minute_slider = Scale(main_window, label="MM", orient=HORIZONTAL, from_=0, to=59, background='SteelBlue3', font=('SegoeUI', 20))

tw_end_minute_slider.grid(row=3, column=2, sticky=W, padx=10, pady=10)

manual_button = Button(main_window, text="Manual Open", bg="Navy", fg="White", command=opensequence, font=('SegoeUI', 26))

manual_button.grid(row=4, column=0, padx=10, pady=10)

log_list = Listbox(main_window, fg='blue', bg='SeaShell3', width=25, height=6, borderwidth=3, font=('SegoeUI', 13))

log_list.grid(row=6, column=0, padx=10, pady=10)

history_label = Label(main_window, bg='LightCyan3', text='Access History', font=('SegoeUI', 15))

history_label.grid(row=5, column=0, padx=10, pady=10, sticky=W)

status_label = Label(main_window, bg='DodgerBlue', text='Current Status: ', font=('SegoeUI', 18))

status_label.grid(row=0, column=2, padx=10, pady=10, sticky=E)

current_status_label = Label(main_window, bg='Salmon', text=status, font=('SegoeUI', 18))

current_status_label.grid(row=0, column=3, padx=10, pady=10, sticky=W)

current_time = Label(main_window, font=('SegoeUI', 15))

current_time.grid(row=0, column=0, columnspan=2, padx=10, pady=10)

def clock():

time_str = datetime.datetime.now().strftime("Current Time: %H:%M:%S")

current_time.config(text=time_str)

if is_within_time_window(datetime.datetime.now().time()):

current_status_label.config(text="Unlocked")

else:

current_status_label.config(text="Locked")

scan()

main_window.after(200, clock)

clock()

mainloop()

Once functioning, the code was tested using the method previously by showing the camera various images of dogs. After it was determined that the code was working as per the engineering requirements, testing was conducted using a new dog, Butch. Video of the results is shown below: