Python Tkinter和动画.FuncAnimation累积延迟和冻结GUI

我正在做一个Tkinter项目，我正在绘制来自不同传感器的实时数据. 我正在使用Tkinter来创建图形用户界面和matplotlib动画.FuncAnimation来创建每给定时间(即1秒)更新的实时绘图.代码是用python3编写的.

只要总点数不多，这项工作就可以了.当点数超过300-400时，系统开始累积延迟、减速并最终冻结. 即，如果我的目标是每1秒进行一次读取，则在开始时，系统每隔1秒和几毫秒返回一次读取.然而，随着时间的推移，它开始以线性趋势增加间隔(请参见下图)

我可以通过创建一个在x轴上具有迭代次数(即读数)、在y轴上具有每次迭代之间的增量时间并每秒更新该图(即使我使用更长的时间间隔，结果也是相同的)的图来重现该问题.

我试着将动画功能放在一个独立的线程中，就像在其他帖子中建议的那样，但完全没有帮助.

如果我没有创建绘图，但我使用生成的数据(即增量时间)来更新标签，则我没有问题，因此它一定与绘图的创建/更新有关.

打开/关闭blit无济于事，不管怎样，我还是宁愿把它关掉.

请参见下面的代码的简短版本，以重现错误和600次迭代后的输出图像.

import tkinter as tk

import matplotlib.pyplot as plt
from matplotlib.backends import backend_tkagg as bk
import matplotlib.animation as animation

import numpy as np

import time
import threading 

class Application(tk.Frame):
    def __init__(self, master=None, **kwargs):
        tk.Frame.__init__(self, master, **kwargs)
       
# =============================================================================
# # Test 1 flags initialization
# =============================================================================
        self.ani = None
        
# =============================================================================
# Canvas frame
# =============================================================================
        self.fig = plt.figure(figsize=(15,5))
        frm = tk.Frame(self)
        frm.pack()

        self.canvas = bk.FigureCanvasTkAgg(self.fig, master=frm)
        self.canvas.get_tk_widget().pack()
        
# =============================================================================
# # Figure initialization
# =============================================================================
        self.ax = self.fig.add_subplot(1,1,1)
        self.ax.plot([],[],'-k', label='delta time')
        self.ax.legend(loc='upper right')
        self.ax.set_xlabel('n of readings [-]')
        self.ax.set_ylabel('time difference between readings [s]')
        
# =============================================================================
# # Start/Quick button
# =============================================================================
        frm4 = tk.Frame(self)
        frm4.pack(side='top', fill='x')

        frm_acquisition = tk.Frame(frm4)
        frm_acquisition.pack()
        self.button= tk.Button(frm_acquisition, text="Start acquisition", command=lambda: self.on_click(), bg='green')
        self.button.grid(column = 0, row=0)

# =============================================================================
# # Methods
# =============================================================================
    def on_click(self):
        '''the button is a start/stop button ''' 
        if self.ani is None:
                self.button.config(text='Quit acquisition', bg='red')
                print('acquisition started')
                return self.start()
        else: 
            self.ani.event_source.stop()
            self.button.config(text='Start acquisition', bg='green')
            print('acquisition stopped')
            self.ani = None
            return
    
    def start(self):
        self.starting_time = time.time()
        self.time = np.array([]) 
        self.ani = animation.FuncAnimation(self.fig, self.animate_thread, interval =1000, blit=False, cache_frame_data=False) #interval in ms     
        self.ani._start()
        return  
    
    # Some post suggested to  put animate() in an indipendent thread, but did not solve the problem
    def animate_thread(self, *args):
        self.w = threading.Thread(target=self.animate)
        self.w.daemon = True  # Daemonize the thread so it exits when the main program finishes
        self.w.start()
        return self.w


    def animate(self, *args):
        self.time = np.append(self.time, time.time()-self.starting_time)
        if len(self.time) > 1:
            self.ax.scatter(len(self.time),self.time[-1]-self.time[-2], c='k')
        # root.update() # another post suggested root.update() but did not help either
        return self.ax,

if __name__ == "__main__":
    root = tk.Tk()
    app=Application(root)
    app.pack()
    root.mainloop()

延时曲线图: