matplotlig speed on plot a gps point on a chart

Question

with the following code I'm trying to add a gps point over a map using a raspberry pi 3:

from gps import *
import matplotlib.pyplot as plt
import time, inspect, os

logPath = "/home/solergy/IBUZ/LOGS/"
inputMapFilePath = logPath + "srcMap/" + "TBM-min.png"
actualOutMapFilePath = logPath + "Map_GPS"

TRX = -12.653          #top right longitude
TRY = 41.8675            #top right latitude
BLX = -12.6332          #bottom left longitude
BLY = 41.8619             #bottom left latitude

gpsd = gps(mode=WATCH_ENABLE|WATCH_NEWSTYLE)

im = plt.imread(inputMapFilePath)

try:
    while True:
        report = gpsd.next() #
        if report['class'] == 'TPV':
            GPStime =  str(getattr(report,'time',''))
            lat     = str(getattr(report,'lat',0.0))
            lon     = str(getattr(report,'lon',0.0))
            speed   =  str(getattr(report,'speed','nan'))

            # Create Image File
            pos_y = float(lat)
            pos_x = -float(lon)

            actualTime = time.strftime("%H-%M-%S", time.localtime())
            plt.text(BLX, BLY, actualTime)
            plt.imshow(im, extent=[BLX, TRX, BLY, TRY], zorder=1)
            plt.scatter(x=[pos_x], y=[pos_y], s=3, c='r', zorder=2)
            cur_axes = plt.gca()
            cur_axes.axes.get_xaxis().set_visible(False)
            cur_axes.axes.get_yaxis().set_visible(False)
            plt.savefig(actualOutMapFilePath, dpi=150, type="png")
            plt.close('all')
            print (actualTime , GPStime)

except (KeyboardInterrupt, SystemExit): #when you press ctrl+c
    f.close()
    f_now.close()

my problem is that I need to update this map at least every second (10Hz is better). The problem is not the gps part but the matplotlib that produce this image. Any idea about how to achieve a higher speed for this code? Antonio

score 0 · Answer 1 · answered May 29 '19 at 08:52

An obvious optimization would be avoid clearing and redrawing everything at every iteraiton, but instead updating the XY values of the dot. See How to update a plot in matplotlib?

I would expect a considerable speedup (especially from avoiding the plt.imshow call), but it is hard to promise anything about performance without access to the actual data (or even hardware).

matplotlig speed on plot a gps point on a chart

1 Answers1