您是否需要以下内容：

rows = [
    ["0", "0", "0", "0", "255", "255", "0", "0", "255"],
    ["255", "255", "0", "0", "0", "255", "0", "0", "255"],
    ["0", "255", "0", "0", "255", "0", "255", "0", "0"],
    ["0", "0", "255", "0", "0", "255", "0", "0", "0"],
]

with open("output.bin", "wb") as f_out:
    for row in rows:
        for field in row:
            f_out.write(int(field).to_bytes())

然后，检查output.bin：

with open("output.bin", "rb") as f_in:
    while True:
        x = f_in.read(9)
        if len(x) == 0:
            break
        print(x)

b'\x00\x00\x00\x00\xff\xff\x00\x00\xff'
b'\xff\xff\x00\x00\x00\xff\x00\x00\xff'
b'\x00\xff\x00\x00\xff\x00\xff\x00\x00'
b'\x00\x00\xff\x00\x00\xff\x00\x00\x00'

感谢Writing integers in binary to file in python向我展示了to_bytes(...)方法，并感谢MyICQ指出了默认值。

这几乎就是所描述的。
我省略了读取变量的输入，这应该是微不足道的。由于输入包含'字符，因此无法将其读取为json。相反，我把它看作是一系列的数字，被一些东西分开。然后应用正则表达式将数字转换为数组。

# Regular expression support
import re

# the input, should be read from file
dirtyinput = "[['0;0;0;0;255;255;0;0;255], ['255;255;0;0;0;255;0;0;255], ['0;255;0;0;255;0;255;0;0], ['0;0;255;0;0;255;0;0;0]]"

# extract numbers
numbers = re.findall(r'\d+', dirtyinput)

# Using function from answer by Zach Young
with open("output.bin", "wb") as f_out:
    for n in numbers:
        f_out.write(int(n).to_bytes(1, 'big'))

# --------- another method, iterating the data (efficient if the data is large)
#
with open("output2.bin", "wb") as f:
    for x in re.finditer(r'\d+', dirtyinput):
        f.write(int(x.group()).to_bytes(1,'big'))

# -------- testing result
# 
with open("output.bin", "rb") as f_in:
    while True:
        x = f_in.read(9)
        if len(x) == 0:
            break
        print(x)

b'\x00\x00\x00\x00\xff\xff\x00\x00\xff'
b'\xff\xff\x00\x00\x00\xff\x00\x00\xff'
b'\x00\xff\x00\x00\xff\x00\xff\x00\x00'
b'\x00\x00\xff\x00\x00\xff\x00\x00\x00'

我得到的结果与上面的答案相同。