Hello, I am interested to know if there's a way to store the coordinates of the detected object only once?
So, when the detection is running it keeps on detecting the object and keeps on appending the coordinate of the same objects again and again(I am using CSV file to write the coordinates). Though there's a slight delta in the coordinates that doesn't matter for now since the object is stationary.
This is what I've now.
for box in y_pred_thresh[0]:
# Transform the predicted bounding boxes for the 300x300 image to the original image dimensions.
xmin = box[2] * 300 / img_width
ymin = box[3] * 300 / img_height
xmax = box[4] * 300 / img_width
ymax = box[5] * 300 / img_height
cv2.rectangle(smallFrame, (int(xmin),int(ymin)),(int(xmax),int(ymax)),(0,250,0), thickness=1,lineType=8)
xcenter=((xmin+xmax)/2) #For center X
ycenter=((ymin+ymax)/2) #For center Y
csv.write(round(xcenter)+ ' , ' + round(ycenter))
csv.write('\n')
csv.close()
cv2.waitKey(0)
return (y_pred_thresh)
Any help is appreciated.
Thanks
* Operating System Ubuntu 16.04
* Which commit of this repository you're on: Latest
* Keras version: 2.2.0
* TensorFlow version: 1.9.0
If someone knows what kind of coordinates do we get? Is it pixel coordinate or something else?
How can I convert it to Robot Coordinate that is in X Y Z?
Hello, I am interested to know if there's a way to store the coordinates of the detected object only once?
So, when the detection is running it keeps on detecting the object and keeps on appending the coordinate of the same objects again and again(I am using CSV file to write the coordinates). Though there's a slight delta in the coordinates that doesn't matter for now since the object is stationary.
This is what I've now.
Any help is appreciated. Thanks
If someone knows what kind of coordinates do we get? Is it pixel coordinate or something else? How can I convert it to Robot Coordinate that is in X Y Z?