On the Fly(OTF)观测产生大量的原始数据，后期需要对原始数据进行预处理，包括数据修正、坏数据剔除、重网格化处理（regrid）等。在国家自然科学基金《基于多波束接收机的OTF观测方法研究与应用》的资助下，在位于德令哈的13.7m毫米波射电望远镜上实现了OTF观测方法，是国内望远镜首次应用OTF观测方法，目前数据预处理采用自己编写程序处理一部分＋应用IRAM 开发的GILDAS软件处理一部分相结合的方式，9个波束的接收机每小时产生16G的数据，两台服务器同时处理需要20分钟，效率较低，GILDAS软件有一些bug和限制，制约着OTF观测设置和数据处理方法，需要改进处理方法，研发处理软件；另外紫金山天文台正在研制100个beam的接收机，届时产出的数据量是目前的10倍多，目前的数据处理方法无法满足100个beam的接收机产生的海量数据的预处理，也需要研发高效的OTF数据处理软件。

On the fly (OTF) observation produces mass original data, which needs to be preprocessed, including data calibration, bad data subtraction, regriding, etc. Supported by NSFC grants <OTF observation method based on multi-beam receiver>, we successfully applied OTF observation method to 13.7 m millimeter telescope at Delingha. This is the first time to use OTF observation method on telescope in our country. We use our primary program together with GILDAS developed by IRAM to preprocess the original data. It produces 16G data by the nine-beam receiver in one hour, which costs 20 min for two workstation to do preprocess. The inefficient preprocessing progress and the bug and limitation of GILDAS software constrain the OTF observation settings and the data processing. Furthermore, the future 100-beam receiver, being developed in purple mountain observatory, will produce 10 times amount of data, which will exceed the ability of data preprocessing method we are using now. Therefore, we need to develop more efficient OTF data processing software.