中文摘要：

针对点云数据局部集中的特点,使用差值预测对点云数据进行预测处理;在预测的同时,根据IEEE-754浮点数标准,简化浮点数的尾数,使用3.5 Byte来表示一个浮点数,以提高压缩效果;然后对预测数据中连续重复的字节使用该字节加该字节重复的次数的方式存储;最后对经过以上处理的数据使用一阶自适应算术编码进行压缩。最终得到的程序在压缩比和内存占用两个方面远优于WinRAR、WinZip压缩软件。

英文摘要：

According to the fact that the Surfel was locally intensive, a predictor was used to predict next Surfel by getting the difference between this Surfel and prior one. At the same time, according to the IEEE-754 floating standards, the mantissa part of a float was reduced from 23 bit to 19 bit. After this, the repeated byte was replaced by this byte and it was repeated times. At the last, the 1 context-based adaptive arithmetic coding was used to further compress the data sets. The result shows that the compression ratio and memory using of the algorithm is greatly better than WinRAR and WinZip.