(紧急求救)linux中 如何高效使用jpeglib 将jpeg图像内容扫描到 buffer中

chenbin200818 chenbin200818 墨衍会员 · AI 创作全网分发 墨衍智能分发 通过墨衍一键同步至各平台 可同步平台 CSDN微信公众号微博知乎掘金百家号博客园抖音小红书 你的文章也可以这样分发 墨衍会员 ¥399起/年,写一次发全网 我也要 + 加入社区
unsigned char* buffer[max];
/*
对buffer分配内存
*/
while (cinfo.output_scanline < jpeg_height)
{
jpeg_read_scanlines(&cinfo,& buffer[i], 1);
i++;
}
对jpeg解压过程中 将扫描到内容读取到buffer中时 发现效率很低
解压一个 宽 4400 高 3000的图片时 while循环 居然用去了10秒,, 有没有办法能高效的执行扫描


求解
...全文

相关推荐

在Linux控制台下使用libjpeg显示JPEG图像在framebuffer上

在Linux控制台下使用libjpeg显示JPEG图像在framebuffer上

liujiu105

V4L2示例代码,测试通过

V4L2示例代码

G_就你不喝可乐是吧

videodog0.31.tar.gz

ideodog is a command line tool, just based on a stock linux distro, with no need of special libraries, or graphic enviroment ( as xlib, imlib). The only special lib is jpeglib, present in most of systems, but you can work without it (see INSTALL). I've running it in an embedded uClinux+busybox installation, with native jpeglib ( compiled for uClibc), and its working fine ! NOTE: Check out MOTION and ALPHABLEND You can use a simple config file, suited for the most common needs, or use the command line options. To change the path, just edit the makefile. The default path is /etc/vd.conf. In the folder called "extra" you will find the gtkpnm, xputjpeg, gtk_cam and a small swap asm function, bgr2rgb. There is a man page. Just check it with man ./videodog.man or move to a proper location. Its not complete yet, so, I didn`t put it in the Makefile. *In a short:* Just grabs a frame from v4l device or set its picture defaults. Useful for scripts, webcam scripts. No dependencies of any external libs. Read LICENSE for more details in copying and distribution. Read INSTALL for details on how to compile/install it. Read CHANGELOG for whats new There is a vd.conf as example configuration, for those who would like to use this feature. Compiling the effects causes no overhead or great executable size increase. Possible outputs: RAW frame (RGB) PNM frame JPEG frame TARGA24 (tga) frame. Double buffer loop capture or single frame with any format. Tested with: CPiA driver (webcam II parport) Bttv driver ( bt 848 & 878 ) ov511 ( webcam III usb ) DLINK cam Notes: - 2.4.x kernel`s CPIA parport driver seems strange for me. Each capture is consuming almost 100% cpu ! - OV511 users: I`ve been receiving reports of VIDIOCMCAPTURE: invalid argument. Don`t know why is it giving this errors, but I can`t test more, because I dont onw such device. Please send reports, if possible. Thanks to : I`ve removed all names from here due to the extreme numer of folks who have been helpful to send bug reports, enhancements, patches, suggestions, or even dropping a line to let me know how videodog was useful to them. Thank you all, and I hope that even with the not-so-constant releases, I can provide not the best, but a solid tool to, hopefully, solve your problems with video capture. Many emails later, and almost 2 years, I`ve been informed that videodog has been running on StrongARM w/o any further mods, that it is the basis of some scientif work, and even being used as webcam capture program. You folks have been very supportive, and this is what keeps me wanting to develop not only videodog, but more open solutions. It is not the best capture tool around, but I think this is why so many people have been using it: its a tool, not a program trying to show it is better than the user, or a ANSI C skill contest. Best regards and thank you again.

chemekom

《模式识别与数据挖掘》配套PPT课件

《模式识别与数据挖掘》配套PPT课件

Bryan Ding

科技成果转化项目进场交易前,评估备案环节需要准备哪些材料?_8.docx

科易网基于AI模型+行业大数据,在技术转移、成果转化、技术经纪、知识产权、产业创新、科技招商等垂直领域的多样化应用场景,推出科技创新领域的AI+工具与数智化解决方案,推动科技创新与产业创新智能化发展。

BIZKEEN
评论 14
收藏 已收藏
分享
返回顶部
chenbin200818
C语言 70,039 成员 243,237 内容