How Wide Is a Point?
I have been reading a lot of banter about accuracy and precision, which brings me to this “point.” Errors in measurements can lead to a wide range of issues in the survey profession. The sum of the potential error equals something I like to call “our error budget." How big is your budget?
First let’s look at the three types of errors that add to that budget.
Gross errors, also known as blunders, can be of any size or nature. They generally occur due to carelessness of the person doing the measuring, like writing down the wrong value or measuring between the wrong points. This error can be reduced or eliminated by using good procedures, but transposing numbers is not uncommon.
Systematic errors are errors that can be mathematically modeled and hence corrected. They are caused by using a different mathematical model than that which exists in the real world. A good example of a systematic error is using the slope distance instead of the horizontal distance. This type of error can be reduced by using the correct and complete model. An incomplete model will add to your error budget and cause the final measurement to be skewed.
Random errors also have an effect on the measurement but have no apparent cause. These errors are the small differences between repeated measurements of the same quantity. They can be reduced or eliminated from our measurements by statistical procedures. One example would be to create a mean value for a measurement to be used for a later measurement.
Now I’m no mathematician or statistical wizard, but I have been around the block a few times in the error buggy. Feedback from others in our profession has always been sharp to criticize a procedure but has also been quick to adopt a procedure once it has been proven beyond a reasonable doubt. This reminds me how some of the latest talks have pointed to the question, “Can scan data produce results that are more accurate than the published specifications of the scanner?” The answer is "Yes," because many points are being used to solve the position of a point. In other words, we can use the power of averages to solve the final position.
The term "survey accuracy" can be confusing, misused and politicized, especially when a layman is involved. Although the fields of mapping and land surveying have benefited greatly from technology, the resounding question of accuracy continues to exist in cadastral surveying based on measurements or law. There are accepted errors in the National Spatial Reference Systems (NSRS), the Continuously Operating Reference Stations (CORS) and the horizontal and vertical control stations maintained by the National Geodetic Survey (NGS). There are errors in the equipment, human errors introduced by the field surveyor, and accepted errors in the technology. But never before have we dissected and expected such precise accuracy in our profession until we knew we could obtain it. What was accepted as standard practice with traditional surveying is no longer good enough with current technology. The days of dead reckoning stood less criticism than .02 foot laser scanner acquired data. Chaining has become obsolete ... which reminds me of a story about an old man that approached a survey crew. The old man asked the crew chief, “Do you still use a chain?" The crew chief replied, “Sure, whenever we get the truck stuck.”
Which makes me ask, just how wide is a point?