MicroPython(OpenMV)如何实现未检测到对象时的代码逃逸路径

你遇到的ArilY is not defined报错和分支逻辑失效，由5个核心问题导致：

• 作用域错误：aril/seed是for循环内部的临时变量，对应色块没被检测到时循环不会执行，变量根本不存在，直接访问就会抛未定义错误；且func_pass/func_fail在全局位置访问循环内的临时变量，必然触发报错
• 执行位置错误：func_orientation()的调用写在了while(True)检测循环外面，程序启动后只会执行1次，不会跟着每帧画面跑检测
• 状态残留：arilY/seedY/rapheY每轮检测前没有重置为空值，会沿用上一帧的检测结果，出现误判
• 语法错误：func_orientation内部的判断逻辑缩进错误、朝向判定的if条件缺失
• 空分支缺失：没有写未检测到目标时的跳过逻辑，检测不到物体时还是会跑朝向校验流程

import sensor, image, time, math
from pyb import UART

sensor.reset()                      # 复位初始化摄像头
sensor.set_pixformat(sensor.RGB565) # 设置图像格式为RGB565
sensor.set_framesize(sensor.QVGA)   # 设置分辨率320x240
sensor.skip_frames(time = 2000)     # 等待设置生效
# 颜色追踪时需关闭自动增益、自动白平衡
#sensor.set_auto_gain(False)
#sensor.set_auto_whitebal(False)

# 色块阈值
threshold_seed = (7,24,-8,4,-3,9)
threshold_aril = (33,76,-14,6,17,69)
threshold_raphe = (36,45,28,43,17,34)

clock = time.clock()                # 帧率统计时钟
uart = UART(3, 9600)                # 初始化串口3，波特率9600

# 全局变量存储检测结果
arilY = None
arilX = None
seedY = None
rapheY = None

def func_pass():
    result = "Pass"
    print(result)
    print("%d\n"%arilX, end='')
    uart.write(result)
    uart.write("%d\n"%arilX)

def func_fail():
    result = "Fail"
    print(result)
    print("%d\n"%arilX, end='')
    uart.write(result)
    uart.write("%d\n"%arilX)

def func_orientation(seedY, arilY):
    # 两个核心色块都检测到才执行判断
    if seedY is not None and arilY is not None:
        # 朝向判定规则：seed在aril上方（Y坐标更小）判定为Pass，可根据实际需求调整判断条件
        if seedY < arilY:
            func_pass()
        else:
            func_fail()

while(True):
    clock.tick()
    img = sensor.snapshot()
    # 每帧检测前重置所有检测结果为空，避免上一帧数据干扰
    arilY = None
    arilX = None
    seedY = None
    rapheY = None

    # 检测seed色块
    for seed in img.find_blobs([threshold_seed], pixels_threshold=200, area_threshold=200, merge=True):
        img.draw_rectangle(seed[0:4])
        img.draw_cross(seed.cx(), seed.cy())
        img.draw_string(seed.x()+2,seed.y()+2,"seed")
        seedY = seed.cy()

    # 检测aril色块
    for aril in img.find_blobs([threshold_aril],pixels_threshold=300,area_threshold=300, merge=True):
        img.draw_rectangle(aril[0:4])
        img.draw_cross(aril.cx(),aril.cy())
        img.draw_string(aril.x()+2,aril.y()+2,"aril")
        arilY = aril.cy()
        arilX = aril.cx()

    # 检测raphe色块
    for raphe in img.find_blobs([threshold_raphe],pixels_threshold=300,area_threshold=300, merge=True):
        img.draw_rectangle(raphe[0:4])
        img.draw_cross(raphe.cx(),raphe.cy())
        img.draw_string(raphe.x()+2,raphe.y()+2,"raphe")
        rapheY = raphe.cy()

    # 空值判断：缺核心检测对象直接打印提示进入下一轮
    if seedY is None or arilY is None:
        print("未检测到目标对象")
        continue

    # 执行朝向判定
    func_orientation(seedY, arilY)

• 每帧开始检测前将所有存储坐标的变量重置为None，避免上一帧旧数据干扰当前帧判断
• 新增空值判断分支：只要seed或aril任意一个核心色块没检测到，直接打印提示后用continue跳过后续判定，进入下一轮循环，从根源避免未定义报错
• 将func_orientation调用移入循环内部，保证每帧画面采集完成后都执行检测逻辑
• 把需要跨作用域访问的aril X坐标单独存为全局变量，解决局部变量跨作用域访问的报错问题
• 补全了朝向判定的条件逻辑、修复了所有缩进错误，可根据实际朝向规则调整seedY < arilY这个判断条件

内容的提问来源于stack exchange，提问作者Hannah Reigi