• 免费好用的星瞳AI云服务上线!简单标注,云端训练,支持OpenMV H7和OpenMV H7 Plus。可以替代edge impulse。 https://forum.singtown.com/topic/9519
  • 我们只解决官方正版的OpenMV的问题(STM32),其他的分支有很多兼容问题,我们无法解决。
  • 如果有产品硬件故障问题,比如无法开机,论坛很难解决。可以直接找售后维修
  • 发帖子之前,请确认看过所有的视频教程,https://singtown.com/learn/ 和所有的上手教程http://book.openmv.cc/
  • 每一个新的提问,单独发一个新帖子
  • 帖子需要目的,你要做什么?
  • 如果涉及代码,需要报错提示全部代码文本,请注意不要贴代码图片
  • 必看:玩转星瞳论坛了解一下图片上传,代码格式等问题。
  • 用edgeimpulse训练好的模型,怎么调用openmv实时的识别概率(即使这个值小于最小置信度也需要)



    • # Edge Impulse - OpenMV Object Detection Example
      
      import sensor, image, time, os, tf, math, uos, gc
      import pyb
      import sensor, image, time
      import json
      from pyb import UART
      
      sensor.reset()                         # Reset and initialize the sensor.
      sensor.set_pixformat(sensor.RGB565)    # Set pixel format to RGB565 (or GRAYSCALE)
      sensor.set_framesize(sensor.QVGA)      # Set frame size to QVGA (320x240)
      sensor.set_windowing((240, 240))       # Set 240x240 window.
      sensor.skip_frames(time=2000)          # Let the camera adjust.
      
      def modified_data(data):
         data = int(data)# 将data转化为整数型变量
         str_data = ''
         if data < 10:                                  # 目标色块的中心点的横坐标、纵坐标或面积的开方<10
             str_data = str_data + '000' + str(data)   # 运用字符串的拼接把色块参数全部转化为长度为四个字符的字符串,如8->“0008”
         elif data >= 10 and data < 100:               # 10<目标色块的中心点的横坐标、纵坐标或面积的开方<100
             str_data = str_data + '00' + str(data)    # 运用字符串的拼接把色块参数全部转化为长度为四个字符的字符串,如88->“0088”
         elif data >=100 and data <1000:               # 100<目标色块的中心点的横坐标、纵坐标或面积的开方<1000
             str_data = str_data + '0' + str(data)     # 运用字符串的拼接把色块参数全部转化为长度为四个字符的字符串,如888->“0888”
         else:                                         # 1000<目标色块的中心点的横坐标、纵坐标或面积的开方
             str_data = str_data + str(data)           # 运用字符串的拼接把色块参数全部转化为长度为四个字符的字符串,如8888->“8888”
         return str_data.encode('utf-8')               # ******将字符串中的每一个字母转化为UTF-8码值(与ASCII码值基本一样)*******
                                                       # ******十进制0对应ASCII码值的48,可进行换算
      
      net = None
      labels = None
      min_confidence = 0.70
      uart = UART(3, 115200)
      
      try:
          # load the model, alloc the model file on the heap if we have at least 64K free after loading
          net = tf.load("trained.tflite", load_to_fb=uos.stat('trained.tflite')[6] > (gc.mem_free() - (64*1024)))
      except Exception as e:
          raise Exception('Failed to load "trained.tflite", did you copy the .tflite and labels.txt file onto the mass-storage device? (' + str(e) + ')')
      
      try:
          labels = [line.rstrip('\n') for line in open("labels.txt")]
      except Exception as e:
          raise Exception('Failed to load "labels.txt", did you copy the .tflite and labels.txt file onto the mass-storage device? (' + str(e) + ')')
      
      colors = [ # Add more colors if you are detecting more than 7 types of classes at once.
          (255,   0,   0),
          (  0, 255,   0),
          (255, 255,   0),
          (  0,   0, 255),
          (255,   0, 255),
          (  0, 255, 255),
          (255, 255, 255),
      ]
      
      clock = time.clock()
      while(True):
          clock.tick()
      
          img = sensor.snapshot()
      
          # detect() returns all objects found in the image (splitted out per class already)
          # we skip class index 0, as that is the background, and then draw circles of the center
          # of our objects
      
          for i, detection_list in enumerate(net.detect(img, thresholds=[(math.ceil(min_confidence * 255), 255)])):
              if (i == 0): continue # background class
              if (len(detection_list) == 0): continue # no detections for this class?
      
              print("********** %s **********" % labels[i])
              for d in detection_list:
                  [x, y, w, h] = d.rect()
                  center_x = math.floor(x + (w / 2))
                  center_y = math.floor(y + (h / 2))
                  print('x %d\ty %d' % (center_x, center_y))
                  img.draw_rectangle(d.rect())
                  print(w*h)
                  print(detection_list[0][4])
                  s = detection_list[0][4]
                  print(s)
                  print(clock.fps(), "fps", end="\n\n")
              if s > 0.75:
                  t = 1
                  x = modified_data(center_x)
                  y = modified_data(center_y)
                  p = modified_data(math.sqrt(w*h))
                  uart.write('st')                                    # 向单片机发送’st’(应该是作为一个发送的起始标志)
                  uart.write(x)                                       # 向单片机发送目标色块的中心点横坐标(经过处理后)
                  uart.write(y)                                       # 向单片机发送目标色块的中心点纵坐标(经过处理后)
                  uart.write(p)
                  time.sleep(0.01)
                  print(t)
              if s < 0.75 and t == 0:
                  uart.write('wz')
                  time.sleep(0.01)
       
      


    • 你把最小置信度参数调低,得到的结果就多了。



    • 还有别的方法吗,因为我们要满足识别成熟果子抓取的条件,置信度不能太低



    • 我没看懂你具体要做什么,你不是想得到包括低置信度的结果吗?传入低置信度参数,哪里不满足你的需求?



    • 就是模型里和最小置信度比较的那个值存在那个变量里?