Programming with MicroPython: KOI

Import KOI Library

Import KOI library to make use of its functions.

from futureKOI import KOI

KOI Basic Operations

Initiate KOI

koi = KOI(tx='P2',rx='P12',id=1)

Initiates the KOI port.

• tx: Tx Pin, use P2 for Robotbit Edu.

• rx: Rx Pin, use P12 for Robotbit Edu.

Set KOI Camera Orientation

# available modes
# 0,2

koi.screen_mode(mode, cmd='K6')

Sets the KOI camera orientation.

• mode: 0 for front, 2 for back

Display text

koi.text(x, y, delay, text, cmd="K4")

Displays a text on the KOI screen.

• x: X coordinate of the text.

• y: Y coordinate of the text.

• delay: Duration of the text in ms.

• text: The string to be displayed.

KOI Screen Capture

koi.screen_save(pic , cmd="K2")

Captures the screen and save to SD card.

• pic: Filename on the SD card in jpg format.

KOI Displays JPG

koi.screen_show(pic, cmd="K1")

Displays a jpg on the screen.

• pic: Filename on the SD card in jpg format.

Face Detection

Load Face Yolo

koi.face_yolo_load(cmd="K30")

Loads the face yolo model.

Single Face Detection

x = koi.face_detect(cmd="K31")

Runs the single face detection.

Obtain Face Coordinates

koi.get_re(cmd="K31")[0]  #x
koi.get_re(cmd="K31")[1]  #y

Returns the X or Y coordinate of the face detected.

Multi Face Detection

koi.face_count(cmd="K32")

Runs the multi face detection.

Obtain Number of Faces

koi.get_re(cmd="K32")[0]

Returns the number of faces detected.

Face Detection Sample Program

from future import *
from futureKOI import KOI

screen.sync = 0
koi = KOI(tx='P2',rx='P12',id=1)
koi.screen_mode(2, cmd='K6')
koi.face_yolo_load(cmd='K30')
while True:
if koi.face_detect(cmd='K31'):
    screen.fill((0, 0, 0))
    screen.text("Face detected:",5,10,1,(255, 255, 255))
    screen.text("X:",5,20,2,(255, 255, 255))
    screen.text((koi.get_re(cmd='K31')[0]),5,40,3,(255, 255, 255))
    screen.text("Y:",5,70,2,(255, 255, 255))
    screen.text((koi.get_re(cmd='K31')[1]),5,90,3,(255, 255, 255))
    screen.refresh()
else:
    screen.fill((0, 0, 0))
    screen.text("No face",5,10,1,(255, 255, 255))
    screen.refresh()
sleep(0.5)

Machine Learning Image Classifier

Initiate Image Classifier

koi.init_cls()

Starts image classifier mode.

Save the Classifier Model

koi.cls_save_model(model,cmd="K43")

Saves the model to the SD Card.

• model: The filename on the SD card.

Load a Classifier Model

koi.cls_load_model(model="abc.json",cmd="K44")

Loads the model from the SD Card.

• model: The filename on the SD card.

Classifier Add Tag

koi.cls_add_tag(tag,cmd="K41")

Add an image tag.

• tag: Name of the tag.

Run Image Classifier

koi.cls_run(cmd="K42")

Returns the tag for the classified object.

KOI Classifier Model Training Sample Program

from future import *
from futureKOI import KOI

items = []
i = 0

items.append('rock')
items.append('paper')
items.append('scissors')
i = 0
koi = KOI(tx='P2',rx='P12',id=1)
koi.init_cls()                                                      # init classifier
koi.screen_mode(2, cmd='K6')    
screen.sync = 0
while True:
screen.fill((0, 0, 0))
if sensor.btnValue("a") and sensor.btnValue("b"):
    koi.cls_save_model(model="model.json",cmd='K43')                # saves the classifier model
    buzzer.melody(1)
else:
    if sensor.btnValue("a"):
    sleep(0.2)
    if not sensor.btnValue("b"):
        koi.cls_add_tag(id=(items[int((i % 3 + 1) - 1)]),cmd='K41') # classifier add tag
        buzzer.melody(4)
    else:
    if sensor.btnValue("b"):
        sleep(0.2)
        buzzer.tone(440,0.2)
        if not sensor.btnValue("a"):
        i += 1
screen.text("Now training:",0,10,1,(255, 255, 255))
screen.text((items[int((i % 3 + 1) - 1)]),0,30,2,(255, 255, 255))
screen.text("Press A to add tag",0,60,1,(255, 255, 255))
screen.text("Press B for next tag",0,80,1,(255, 255, 255))
screen.text("Press A+B to save",0,100,1,(255, 255, 255))
screen.refresh()

KOI Image Classification Sample Program

from future import *
from futureKOI import KOI

koi = KOI(tx='P2',rx='P12',id=1)
koi.screen_mode(2, cmd='K6')
koi.init_cls()
koi.cls_load_model(model="model.json",cmd='K44')                    # loads the classifier model
while True:
if sensor.btnValue("a"):
    screen.clear()
    screen.text((koi.cls_run(cmd='K42')),5,10,2,(255, 255, 255))    # displays the classified tag
    screen.refresh()

Colour Blob and Line Tracking

Calibrate a Colour to Track

koi.color_cali(name ,cmd="K16")

Calibrates a colour to track.

• name: The name for this colour.

Tracks a Colour Blob

koi.color_tracking(name="name", cmd="K15")
# returns [cx,cy,w,h,name]

Returns a list containing the values of the colour blob.

• name: the name of the colour to track.

Obtain Colour Blob Values

koi.get_re(cmd="K15")[0] #cx
koi.get_re(cmd="K15")[1] #cy
koi.get_re(cmd="K15")[2] #w
koi.get_re(cmd="K15")[3] #h

Returns the specific value of the colour blob.

Blob Detection Sample Program

from future import *
from futureKOI import KOI

koi.screen_mode(2, cmd='K6')
koi = KOI(tx='P2',rx='P12',id=1)
while True:
if sensor.btnValue("a"):
    koi.color_cali(name="red" ,cmd='K16')
    sleep(0.3)
if sensor.btnValue("b"):
    if koi.color_tracking(name="red", cmd='K15'):
    screen.clear()
    screen.text((koi.get_re(cmd='K15')[0]),5,10,1,(255, 255, 255))
    screen.text((koi.get_re(cmd='K15')[1]),5,20,1,(255, 255, 255))
    screen.text((koi.get_re(cmd='K15')[2]),5,30,1,(255, 255, 255))
    screen.text((koi.get_re(cmd='K15')[3]),5,40,1,(255, 255, 255))
    screen.refresh()
    sleep(0.3)

Tracks a line

koi.line_tracking(name ,cmd="K12")
# returns [x1, y1, x2, y2, name]

Returns a list containing the values of the line.

• name: the name of the colour to track.

Obtain Line Values

koi.get_re(cmd="K12")[0] #x1
koi.get_re(cmd="K12")[1] #y1
koi.get_re(cmd="K12")[2] #x2
koi.get_re(cmd="K12")[3] #y2

Returns the specific value of the line.

Line Detection Sample Program

from future import *
from futureKOI import KOI

koi.screen_mode(2, cmd='K6')
koi = KOI(tx='P2',rx='P12',id=1)
while True:
if sensor.btnValue("a"):
    koi.color_cali(name="red" ,cmd='K16')
    sleep(0.3)
if sensor.btnValue("b"):
    if koi.line_tracking(name="red" ,cmd='K12'):
    screen.clear()
    screen.text((koi.get_re(cmd='K12')[0]),5,10,1,(255, 255, 255))
    screen.text((koi.get_re(cmd='K12')[1]),5,20,1,(255, 255, 255))
    screen.text((koi.get_re(cmd='K12')[2]),5,30,1,(255, 255, 255))
    screen.text((koi.get_re(cmd='K12')[3]),5,40,1,(255, 255, 255))
    screen.refresh()
    sleep(0.3)

Shape Detection

Circle Detection

koi.circle_detect(threshold, cmd="K10")

Runs the circle detection.

• threshold: The threshold for circle, typical value is about 2000, adjust to achieve best results.

Obtain Circle Values

koi.get_re(cmd="K10")[0] #cx
koi.get_re(cmd="K10")[1] #cy
koi.get_re(cmd="K10")[2] #r

Returns the specific value of the circle.

Rectangle Detection

 koi.rectangle_detect(th=4000,cmd="K11")

Runs the rectangle detection.

• threshold: The threshold for rectangle, typical value is about 1000, adjust to achieve best results.

Obtain Rectangle Values

koi.get_re(cmd="K11")[0]
koi.get_re(cmd="K11")[1]
koi.get_re(cmd="K11")[2]
koi.get_re(cmd="K11")[3]

Returns the specific value of the rectangle.

Shape Detection Sample Program

from futureKOI import KOI
from future import *

koi.screen_mode(2, cmd='K6')
koi = KOI(tx='P2',rx='P12',id=1)
while True:
if sensor.btnValue("a"):
    if koi.circle_detect(th=4000, cmd='K10'):
    screen.clear()
    screen.text((koi.get_re(cmd='K10')[0]),5,10,1,(255, 255, 255))
    screen.text((koi.get_re(cmd='K10')[0]),5,20,1,(255, 255, 255))
    screen.text((koi.get_re(cmd='K10')[0]),5,30,1,(255, 255, 255))
    screen.refresh()
    sleep(0.3)
if sensor.btnValue("b"):
    if koi.rectangle_detect(th=4000,cmd='K11'):
    screen.clear()
    screen.text((koi.get_re(cmd='K11')[0]),5,10,1,(255, 255, 255))
    screen.text((koi.get_re(cmd='K11')[1]),5,20,1,(255, 255, 255))
    screen.text((koi.get_re(cmd='K11')[2]),5,30,1,(255, 255, 255))
    screen.text((koi.get_re(cmd='K11')[3]),5,40,1,(255, 255, 255))
    screen.refresh()
    sleep(0.3)

Code Scanner

QR Code Scanner

koi.scan_qrcode(cmd="K20")

Returns the QR Code data.

Barcode Scanner

koi.scan_barcode(cmd="K22")

Returns the Barcode data.

Apriltag Scanner

koi.scan_Apriltag(cmd="K23")

Runs the scanner for Apriltag.

Obtains Apriltag data

koi.get_re(cmd="K23")[0] #id
koi.get_re(cmd="K23")[1] #cx
koi.get_re(cmd="K23")[2] #cy
koi.get_re(cmd="K23")[3] #w
koi.get_re(cmd="K23")[4] #h

Returns the specific data of the Apriltag.

Code Scanner Sample Program

from future import *
from futureKOI import KOI

koi = KOI(tx='P2',rx='P12',id=1)
koi.screen_mode(2, cmd='K6')
while True:
if sensor.btnValue("a"):
    screen.clear()
    screen.text((koi.scan_qrcode(cmd='K20')),5,10,2,(255, 255, 255))
if sensor.btnValue("b"):
    screen.clear()
    screen.text((koi.scan_barcode(cmd='K22')),5,10,2,(255, 255, 255))

Wifi Connection

Connect to Wifi

koi.connect_wifi(router ,pwd ,cmd="K50")

Connects to a Wifi network.

• router: The SSID.

• pwd: The password.

Displays IP Address

koi.get_ip(cmd="K54")

Displays the IP Address.

BaiduAI

BaiduAI Face Recognition

koi.baiduAI_face_detect(cmd="K75")
# returns [face token, age, sex, mask, emotion]

Returns a list of values from the BaiduAI Face Recognition.

BaiduAI Face Token

koi.get_re(cmd="K75")[0]

Returns the face token.

BaiduAI Add Face to Group

koi.baiduAI_face_add(face_token="token" ,groupName="group" ,faceName="name" ,cmd="K76")

Adds the face token to a group.

• face_token: The token to be added.

• groupName: The group name.

• faceName: The name of the person.

BaiduAI Search Face in Group

koi.baiduAI_face_search(face_token="token" ,groupName="group" ,cmd="K77")[0]

Searches the face in a group and returns the person name.

• face_token: The token to be searched.

• groupName: The group name.

BaiduAi Text to Speech

koi.baiduAI_tts(text ,cmd="K78")

Generates a speech using BaiduAI.

-text: The text to be spoken.

Wifi and Baidu TTS Sample Program

from future import *
from futureKOI import KOI

koi = KOI(tx='P2',rx='P12',id=1)
koi.screen_mode(2, cmd='K6')
koi.connect_wifi(router="apname" ,pwd="password" ,cmd='K50')
while True:
if sensor.btnValue("a"):
    koi.baiduAI_tts(txt='"hello"' ,cmd='K78')
    sleep(0.2)
if sensor.btnValue("b"):
    screen.clear()
    screen.text((koi.get_ip(cmd='K54')),5,10,1,(255, 255, 255))
    screen.refresh()
    sleep(0.2)

Audio

Record Wav File

koi.audio_record(name)

Record and saves the wav file to the SD card.

• name: The file name in wav format.

Play Wav File

koi.audio_play(name)

Plays a wav file from the SD card.

• name: The file name in wav format.

Calibrate Ambient Noise

koi.audio_noisetap()

Calibrates the ambient noise for speech recognition.

Speech Recognition Add Tag

koi.speech_add_tag(tag)

Adds a speech tag.

• tag: The tag name.

Get Speech Recognition Tag

koi.speech_run(cmd="K65")

Runs the speech recognition and returns the tag.

Save Speech Model

koi.speech_save_model(file)

Saves the speech model to the SD card.

• file: The filename.

Load Speech Model

koi.speech_load_model(file)

Loads a speech model.

• file: The filename.

Speech Recognition Training Sample Program

from future import *
from futureKOI import KOI

items = []
i = 0


items.append('rock')
items.append('paper')
items.append('scissors')
i = 0
koi = KOI(tx='P2',rx='P12',id=1)
koi.audio_noisetap()
koi.screen_mode(2, cmd='K6')
screen.sync = 0
while True:
screen.fill((0, 0, 0))
if sensor.btnValue("a") and sensor.btnValue("b"):
    buzzer.melody(1)
    koi.speech_save_model("speech.json")
else:
    if sensor.btnValue("a"):
    sleep(0.2)
    if not sensor.btnValue("b"):
        koi.speech_add_tag((items[int((i % 3 + 1) - 1)]))
    else:
    if sensor.btnValue("b"):
        sleep(0.2)
        buzzer.tone(440,0.2)
        if not sensor.btnValue("a"):
        i += 1
screen.text("Now training:",0,10,1,(255, 255, 255))
screen.text((items[int((i % 3 + 1) - 1)]),0,30,2,(255, 255, 255))
screen.text("Press A to add tag",0,60,1,(255, 255, 255))
screen.text("Press B for next tag",0,80,1,(255, 255, 255))
screen.text("Press A+B to save",0,100,1,(255, 255, 255))
screen.refresh()

Speech Recognition Sample Program

from future import *
from futureKOI import KOI


koi = KOI(tx='P2',rx='P12',id=1)
koi.audio_noisetap()
koi.speech_load_model("speech.json")
while True:
if sensor.btnValue("a"):
    screen.clear()
    screen.text((koi.speech_run(cmd='K65')),5,10,2,(255, 255, 255))
    screen.refresh()

Miscellaneous

Reset KOI

koi.reset(cmd='k99')

Resets the KOI.

Stop Classifier and Face Yolo

koi.stop_kpu(cmd='k98')

Stops the KPU from running Classifier and Face Yolo.