这条新命令允许您在对基表的一次扫描中收集统计数据，并构建索引，而不像这些特性的以前实现那样需要进行两遍扫描。
This new command allows you to collect statistics and build an index in a single pass of the base table, instead of two passes with the previous implementation of these features.
我的方法是，从左到右，从上到下对谜板进行扫描(两遍，这是为了得到垂直方向和水平方向的词)。
Instead, I scan the board from left to right and top to bottom (twice, to get both horizontal and vertical words).
整个识别过程只需扫描一遍图像和两遍连通域队列。
The recognition process needed scanning the whole images once and connected domain queue twice.
应用推荐