Meaning of stripe in image

image stride（ Image line span ） That is, the control occupied by each row of pixels in memory , For memory alignment or other reasons , The space occupied by each row of pixels in memory is not the width of the image .

As shown in the figure , For image data , Its pixel storage is row by row , However, different compression methods may be used for different image formats , These compressed data also need to be stored in the image , Easy to read and write . In the diagram above , In the picture height Is the height of the image , The width of the image is not the size of a row of data , But only part of the data .stride yes width as well as padding And .width Even if the frame of the image , and padding Is used to store information related to image format （ For example, an image of the same size ,8bit And 32bit Its memory alignment is different ,stride It's different ）.pitch and stride similar , It's just pitch For pixels ,stride For bytes . The following figure shows how images are stored

Span -stride

We all know the of computers now cpu All are 32 Bits or 64 Bit cpu, They read at least... At a time 4、8 Bytes , If less than these , Instead, do some extra work , It will take longer . All have a concept called memory alignment , Set the length of the structure to 4、8 Multiple .

The span appears for the same reason . Because the image operation is usually operated on a line by line basis , If all the data of the image are closely arranged , Then there will be very many reads of misaligned memory . Will affect efficiency . And image processing is a race against the clock operation , Therefore, in order to improve the performance, the concept of span is introduced .
Span refers to the length of the storage space occupied by a row of image data in the image , It is a memory alignment length greater than or equal to the image width . In this way, the memory can be aligned every time the data is read based on the behavior benchmark , Although there may be a little memory waste , But today, with plenty of memory, it doesn't matter .

https://www.jianshu.com/p/68e05ad85490

SenLin - You know