借助百度api,用python实现.

思路

  1. 使用python录音并保存为文件
  2. 调用api,得到识别结果
  3. 根据结果打开对应url

只贴了关键代码,完整代码请移步:https://github.com/missyoursmile/Command-a-computer-with-sound

#定义命令字典(不知道为何百度会识别成这样,牛客识别不出来,用了部分识别)
url_dict = {
    u'ｂａｉｄｕ': "https://www.baidu.com",
    u'牛': "https://www.nowcoder.com"
}
#录音并保存到文件(在另外一个文件中)
wavfile = record_wave()

#获取token
access_token = requests.post(auth_url).json()['access_token']
res = None
with open(wavfile,'rb') as f:
    r = requests.post(api_url+access_token, headers={'Content-Type': "audio/wav;rate=8000"}, data=f)

    res = r.json()
    if res['err_no'] != 0: #识别错误
        res = None

if not res:
    print "识别错误"
    exit()
open_url = None
for k in url_dict:
    if k in res['result'][0]:
        open_url = url_dict[k]

print "识别结果: ",res['result'][0]

if not open_url:
    print "没有对应的命令."
    exit()
#调用系统命令,打开浏览器
os.system("explorer "+ open_url)

注意