怎样拍刀口阴影

我的口径310焦距770抛物面就要结束了。在平面镜前刀口阴影基本无误差，光栅很平直。今天弄来一个数码相机想拍下刀口阴影请大家指导。可是将光圈放到2。8（最大），快门15秒（最长时间）仍无法拍下刀口阴影。请教各位高手有谁拍过刀口阴影指导指导。

懒得打字了，转一个：
http://foucault.sourceforge.net/instructions.html

Taking Images
To get meaningful results from image analysis, you need to take some precautions to ensure a good quality of images without large systematic effects. Here are some tips to ensure an accurate measurement:

Place your camera as close to the knife as possible. You will be taking a sequence of measurements with the knife at different offsets, so for initial setup place the knife at the rearmost location and adjust the camera lens so it almost touches the knife.
Fix the aperture, duration of exposure, focus and white balance of your camera throughout an entire measurement. You might be able to get meaningful results with a fully automatic ("idiot") camera, but it is certainly not worth the doubt if you can buy or borrow a camera with a "full manual" mode.
Try to keep the camera steady throughout a measurement. If you nudge it so the image of the mirror shifts a few pixels in the images that's OK, it can be taken care of in the program. If, on the other hand, you inadvertently change the zoom or the distance of the objective from the mirror, discard the measurement and start again.
Set a long enough exposure to get good contrast and to escape low-level random noise inherent in CCD devices, but not so long that you saturate pixels. To give you an idea, a half-second exposure seems to work well with a typical consumer camera and a three-candle green LED from Radio Shack.
Minimize background light, especially reflected sunlight if working during the day. On a partly cloudy day variations in sunlight intensity can render a measurement useless. Of course, the ideal setting is working in complete darkness with a black screen behind the mirror, but you don't have to go overboard to get accurate readings.
Use a large LED or a long slit. When the knife is far from the radius of curvature you'll be taking images not of a single point but rather of a relatively large section of the light source, as different points get reflected by different areas on the mirror.
Use "optical" and "digital" zoom and smaller image sizes so you can fit more and finer-grained measurements on your hard disk. If your camera won't zoom that close, you can also crop useless regions of images in an image editing program to save space. These image collections tend to get quite large over time...

Image format doesn't seem to make any practical difference. Lossy formats such as JPEG tend to lose information in an unbiased manner, so given a choice by your camera use any format that lets you store enough images to complete an entire sequence without having to download and erase images in the middle of a measurement. The program currently supports JPEG, PNG and GIF formats.