You can set the encoding of the file you are reading from, so

df = sqlContext
    .read
    .option("encoding", "yourEncoding")//replace yourEncoding with whatever your input file encoding is, I asume latin-1?
    .csv("file.csv", header=False, schema = schema)

more on available options here: https://spark.apache.org/docs/latest/sql-data-sources-csv.html#data-source-option