图：CMU3

相比CMU2+，CMU3最大的不同是开源了。另外还支持了闪存卡的直接读写，也就是可以将图像及时的存入CF卡里，供以后分析。另外CMU3本身是不带图像处理的，只提供了基本的压缩，直方图等功能，需要高级的图像处理功能，需要用户自己载入相应的处理程序。比如人脸检测，CMU2+模拟器等。CMU2+模拟器可以让CMU3获得和CMU2+完全一样的图像处理能力，包括颜色识别，追踪，运动检测等。

CMU3的详细技术参看下表：

• Fully open source and programmable using GCC
  CIF resolution (352x288) RGB color sensor
  Open source development environment for Windows and Linux
  MMC/SD flash slot with FAT16 driver support
  Four-port servo controller
  Image processing rate of 26 frames per second
  Lua light-weight language interpreter for rapid prototyping
  Software JPEG compression
  Basic image manipulation library
  Arbitrary image clipping
  Image downsampling
  Threshold and convolution functions
  RGB, YCrCb and HSV color space
  CMUcam2 emulation
  User defined color blobs
  Frame differencing
  Mean and variance data collection
  Raw images dumps over serial
  Histogram generation
  B/W analog video output (PAL or NTSC)
  FIFO image buffer for multiple pass hi-res image processing
  Wireless Mote networking interface (802.15.4)
  Tmote Sky / Telos Connection
  virtual-cam for prototyping on the PC
  CMUcam3-Frame-Grabber for viewing images on the PC
  

相关链接：

详细资料和下载：http://cmucam.org/wiki/Downloads