MyCobot 抓取识别颜色问题：识别黄色和红色可以，其他颜色不行
MyCobot 抓取识别颜色问题：识别黄色和红色可以，其他颜色不行
这段代码中，self.color 用于记录检测到的方块的颜色，并在后面的夹取动作中使用。代码在 color_detect() 函数中使用 self.color 记录检测到的方块颜色。
当检测到黄色和红色时可以识别，而检测其他颜色不能识别，可能是由于 HSV 颜色空间的设定不准确，需要根据实际情况进行调整。
以下是代码片段和可能的解决方案：
#encoding: UTF-8
#!/usr/bin/env python2
import cv2
import os
import numpy as np
import time
# import rospy
from pymycobot.mycobot import MyCobot
from opencv_yolo import yolo
from VideoCapture import FastVideoCapture
import math
import basic
import argparse
parser = argparse.ArgumentParser(description='manual to this script')
parser.add_argument("--debug", type=bool, default="True")
args = parser.parse_args()
# < ...................................parameter...................................
ratio = 0.193
x_bias = -3
y_bias = 30
# increase height_bias to move higher, or decrease height_bias to move lower
height_bias = 148
grab_direct = 'front'
# coords_ready = [190, -30, 245, -175, -0, -45]
coords_ready = [0, 0, 0, 0, 0, 0]
GRAB_MOVE_SPEED = 50
debug = True# show image and waitkey
IMG_SIZE = 640
done = False
cap_num = 3
usb_dev = "/dev/arm"
baudrate = 115200
# < ...................................parameter...................................
coords = coords_ready
# angles = [-90, -40, -10, -35, 0, -45]
class Detect_marker(object):
def __init__(self):
super(Detect_marker, self).__init__()
self.mc = MyCobot(usb_dev, baudrate)
self.mc.power_on()
self.yolo = yolo()
# choose place to set cube
self.color = 0
# parameters to calculate camera clipping parameters
self.x1 = self.x2 = self.y1 = self.y2 = 0
# set color HSV
self.HSV = {
'yellow': [np.array([20, 43, 46]), np.array([26, 255, 255])],
'red': [np.array([0, 43, 46]), np.array([10, 255, 255])],
'green': [np.array([50, 43, 46]), np.array([65, 255, 255])],
'blue': [np.array([100, 43, 46]), np.array([124, 255, 255])],
'purple': [np.array([125, 43, 46]), np.array([155, 255, 255])],
}
# use to calculate coord between cube and mycobot
self.sum_x1 = self.sum_x2 = self.sum_y2 = self.sum_y1 = 0
# The coordinates of the cube relative to the mycobot
self.c_x, self.c_y = IMG_SIZE/2, IMG_SIZE/2
# The ratio of pixels to actual values
self.ratio = ratio
# detect cube color
def color_detect(self, img):
# set the arrangement of color'HSV
x = y = 0
for mycolor, item in self.HSV.items():
redLower = np.array(item[0])
redUpper = np.array(item[1])
# transfrom the img to model of gray
hsv = cv2.cvtColor(img, cv2.COLOR_BGR2HSV)
# wipe off all color expect color in range
mask = cv2.inRange(hsv, item[0], item[1])
# a etching operation on a picture to remove edge roughness
erosion = cv2.erode(mask, np.ones((1, 1), np.uint8), iterations=2)
# the image for expansion operation, its role is to deepen the color depth in the picture
dilation = cv2.dilate(erosion, np.ones(
(1, 1), np.uint8), iterations=2)
# adds pixels to the image
target = cv2.bitwise_and(img, img, mask=dilation)
# the filtered image is transformed into a binary image and placed in binary
ret, binary = cv2.threshold(dilation, 127, 255, cv2.THRESH_BINARY)
# get the contour coordinates of the image, where contours is the coordinate value, here only the contour is detected
contours, hierarchy = cv2.findContours(
dilation, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
if len(contours) > 0:
# do something about misidentification
boxes = [
box
for box in [cv2.boundingRect(c) for c in contours]
if 110 < min(box[2], box[3]) and max(box[2], box[3]) < 170
# if min(img.shape[0], img.shape[1]) / 10
# < min(box[2], box[3])
# < min(img.shape[0], img.shape[1]) / 1
]
print(boxes)
if boxes:
for box in boxes:
# print(box)
x, y, w, h = box
# if abs(w-h)>15:
# return None
# find the largest object that fits the requirements
c = max(contours, key=cv2.contourArea)
# get the lower left and upper right points of the positioning object
x, y, w, h = cv2.boundingRect(c)
print(x, y, w, h)
# locate the target by drawing rectangle
# cv2.rectangle(img, (x, y), (x+w, y+h), (153, 153, 0), 2)
# cv2.rectangle(img, (x, y), (x+130, y+130), (255, 0, 255), 2)
# calculate the rectangle center
x, y = (x*2+w)/2, (y*2+h)/2
# calculate the real coordinates of mycobot relative to the target
if mycolor == 'yellow':
self.color = 1
elif mycolor == 'red':
self.color = 0
elif mycolor == 'green':
self.color = 3
elif mycolor == 'blue':
self.color = 2
elif mycolor == 'purple':
self.color = 4
# print("color:", self.color)
else:
self.color = 1
if abs(x) + abs(y) > 0:
return x, y
else:
return None
# Grasping motion
def move(self, x, y):
global height_bias, done
coords_target = [coords[0]+int(x), coords[1]+int(y), height_bias, coords[3], coords[4], coords[5]]
basic.move_to_target_coords(coords_target, GRAB_MOVE_SPEED)
basic.grap(True)
angles = [0, 0, 0, 0, 0, 0]
self.mc.send_angles(angles,30)
time.sleep(3)
done = True
print("Done")
self.mc.set_color(0,255,0)#green, arm is free
# init mycobot
def init_mycobot(self):
# basic.grap(False)
# time.sleep(1)
# basic.move_to_target_coords([190, -30, 245, -175, -0, -45],GRAB_MOVE_SPEED)
self.mc.send_angles([-90, -40, -10, -35, 0, -45],30)
def get_position(self, x, y):
wx = wy = 0
if grab_direct == 'front':
wx = (self.c_y - y) * self.ratio
wy = (self.c_x - x) * self.ratio
elif grab_direct == 'right':
wx = (self.c_x - x) * self.ratio
wy = (y - self.c_y) * self.ratio
return wx, wy
def transform_frame(self, frame):
frame, ratio, (dw, dh) = self.yolo.letterbox(frame, (IMG_SIZE, IMG_SIZE))
return frame
def run(self):
self.mc.set_color(255, 0, 0)#blue, arm is busy
self.init_mycobot()
def show_image(self, img):
# print(args.debug)
if debug and args.debug:
cv2.imshow("figure", img)
cv2.waitKey(50)
if __name__ == "__main__":
detect = Detect_marker()
detect.run()
cap = FastVideoCapture(cap_num)
while True:
frame = cap.read()
#让图像顺时针旋转90度
frame = cv2.rotate(frame, cv2.ROTATE_90_CLOCKWISE)
# frame = detect.transform_frame(frame)
detect.show_image(frame)
# while cv2.waitKey(100) < 0 and not done:
# frame = cap.read()
# # cv2.imshow('pp',frame)
# # deal img
# frame = detect.transform_frame(frame)
# # get detect result
# detect_result = detect.color_detect(frame)
# detect.show_image(frame)
# # cv2.rectangle(frame, (x-10, y-10), (x+10, y+10), (0,255,0),2)
# if detect_result is None:
# continue
# else:
# x, y = detect_result
# # calculate real coord between cube and mycobot, unit mm
# real_x, real_y = detect.get_position(x, y)
# cv2.rectangle(frame, (x-10, y-10), (x+10, y+10), (0,255,0),2)
# print(real_x, real_y)
# detect.move(real_x, real_y)
# coords_now = basic.get_coords()
# if len(coords_now) == 6:
# coords = coords_now
# detect.move(real_x + x_bias, real_y + y_bias)
# cap.close()
可能的解决方案：
-
调整 HSV 颜色空间范围：
- 使用颜色选择工具 (例如 OpenCV 的
cv2.inRange()函数) 确定每个颜色的准确 HSV 范围。 - 可以通过调整
self.HSV字典中的颜色范围来提高识别精度。
- 使用颜色选择工具 (例如 OpenCV 的
-
使用更精确的颜色识别算法：
- 探索更先进的颜色识别算法，例如 K-Means 聚类、颜色直方图分析等，以提高识别精度。
-
使用机器学习模型：
- 训练一个机器学习模型来识别不同的颜色。这需要收集大量不同颜色的图像数据，并使用训练好的模型来识别颜色。
建议：
- 使用颜色选择工具仔细调整 HSV 范围，以确保每个颜色都能被准确识别。
- 测试不同颜色识别算法，选择最适合当前应用场景的算法。
- 如果需要更高精度的颜色识别，可以考虑使用机器学习模型。
希望以上内容能帮到您！如果您还有其他问题，请随时提问。
原文地址: https://www.cveoy.top/t/topic/gvX0 著作权归作者所有。请勿转载和采集!