资源描述
Section 9 CONCEPTS OF REMOTE SENSINGINTRODUCTIONRemote sensing is the science and art of obtaining information about an object, area, or phenomenon through the analysis of data acquired by a device that is not in contact with the object, area, or phenomenon under investigation. As you read these words you are employing remote sensing. Your eyes are acting as sensors that respond to the light reflected from this page. The data your eyes acquire are impulses corresponding to the amount of light reflected from the dark and light areas on the page. These data are analyzed, or interpreted, in your mental computer to enable you to explain the dark areas on the page as a collection of letters forming words. Beyond this, you recognize that the words form sentences, and interpret the information that the sentences convey.In many respects, remote sensing can be thought of as a reading process. Using various sensors we remotely collect data that may be analyzed to obtain information about the objects, areas, or phenomena being investigated. The remotely collected data can be of many forms, including variations in force distributions, acoustic wave distributions, or electromagnetic energy distributions. For example, a gravity meter acquires data on variations in the distribution of the force of gravity. Sonar, like a bats navigation system, obtains data on variations in acoustic wave distributions. Our eyes acquire data on variations in electromagnetic energy distributionsThis book is about electromagnetic energy sensors that are currently being operated from airborne and spaceborne platforms to assist in inventorying, mapping, and monitoring earth resources. These sensors acquire data on the way various earth surface features emit and reflect electromagnetic energy and these data are analyzed to provide information about the resources under investigation. Figure 1 schematically illustrates the generalized processes and elements involved in electromagnetic remote sensing of earth resources. The two basic processes involved are data acquisition and data analysis. The elements of the data acquisition process are energy sources (a), propagation of energy through the atmosphere (b), energy interactions with earth surface features (c), re-transmission of energy through the atmosphere (d), airborne and/or spaceborne sensors (e), resulting in the generation of sensor data in pictorial and/or digital form (f). In short, we use sensors to record variations in the way earth surface features reflect and emit electromagnetic energy. The data analysis process (g) involves examining the data using various viewing and interpretation devices to analyze pictorial data, and/or a computer to analyze digital sensor data. Reference data about the resources being studied (such as soils maps, crop statistics, or field-check data) are used when and where available to assist in the data analysis. With the aid of the reference data, the analyst extracts information about the type, extent, location, and condition of the various resources over which the sensor data were collected. This information is then compiled (h), generally in the form of hard copy maps and tables, or as computer files that can be merged with other layers of information in a Geographic Information System (GIS). Finally, the information is presented to users (i) who apply it to their decision-making process. In the remainder of this chapter, we discuss the basic principles underlying the remote sensing process. We begin with the fundamentals of electromagnetic energy, then consider how the energy interacts with earth surface features. We also treat the role that reference data play in the data analysis procedure. These basics will permit us to conceptualize an ideal remote sensing system. With that as a framework, we consider the limitations encountered in real remote sensing systems. At the end of this discussion, the reader should have a grasp of the general concepts and foundations of remote sensing. DATA ACQUISITION AND INTERPRETATIONUp to this point, we have discussed the principal sources of electromagnetic energy, the propagation of this energy through the atmosphere, and the interaction of this energy with earth surface features. Combined, these Factors result in energy signals from which we wish to extract information. We now consider the procedures by which these signals are detected, recorded, and interpreted. The detection of electromagnetic energy can be performed either photographically or electronically. The process of photography uses chemical reactions on the surface of a light sensitive film to detect energy variations within a scene. Photographic systems offer many advantages: they arc relatively simple and inexpensive and provide a high degree of spatial detail and geometric integrity. Electronic sensors generate an electrical signal that corresponds to the energy variations in the original scene. A familiar example of an electronic sensor is a video camera. Although considerably more complex and expensive than photographic systems, electronic sensors offer the advantages of a broader spectral range of sensitivity, improved calibration potential, and the ability to electronically transmit data. By developing a photograph, we obtain a record of its detected signals. Thus, the film acts as both the detecting and the recording medium. Electronic sensor signals are generally recorded onto magnetic tape. Subsequently, the signals may be converted to an image form by photographing a TV-like screen display of the data, or by using a specialized film recorder. In these cases, photographic film is used only as a recording medium. In remote sensing, tile term photograph is reserved exclusively for images that were detected as well as recorded on film. The more generic term image is used for any pictorial representation of image data. Thus, a pictorial record from a thermal scanner (an electronic sensor) would be called a thermal image, not a thermal photograph, because film would not be the original detection mechanism for the image. Because the term image relates to any pictorial product, all photographs are images. Not all images, however, are photographs. We can see that the data interpretation aspects of remote sensing can involve analysis of pictorial (image) and/or digital data. Visual interpretation of pictorial image data has long been the workhorse of remote sensing. Visual techniques make use of the excellent ability of the human mind to qualitatively evaluate spatial patterns in a scene. The ability to make subjective judgments based on selective scene elements is essential in many interpretation efforts. Visual interpretation techniques have certain disadvantages, however, in that they may require extensive training and are labor intensive. In addition, spectral characteristics are not always fully evaluated iii visual interpretation efforts. This is partly because of the limited ability of the eye to discern tonal values on an image and the difficulty for an interpreter to simultaneously analyze numerous spectral images. In applications where spectral patterns are highly informative, it is therefore preferable to analyze digital, rather than pictorial, image data. The basic character of digital image data is illustrated in Figure 1.11. Though the image shown in (a) appears to be a continuous tone photograph, it is actually composed of a two-dimensional array of discrete picture elements or pixels. The intensity of each pixel corresponds to the average brightness or radiance measured electronically over the ground area corresponding to each pixel. A total of 320 rows and 480 columns of pixels are shown in Figure 1.1la. Whereas the individual pixels are virtually impossible to discern in (a), they are readily observable in the enlargements shown in (b) and (c). These enlargements correspond to subareas located in the vicinity of the in (a). A 19 row x 27 column enlargement is shown in (b) and a 10 row x 15 column enlargement is included in (c). Part (d) shows the individual digital number (DN) corresponding to the average radiance measured in each pixel shown in (c). These values are simply positive integers that result from quantizing the original electrical signal from the sensor into positive integer values using a process called analog-to-digital (A-to-D) signal conversion. Typically, the DNs constituting a digital image are recorded over such numerical ranges as 0 to 63, 0 to 127, 0 to 255, 0 to 511, or 0 to 1023. These ranges represent the set of integers that can be recorded using 6-, 7-, S-, 9-, and 10-bit binary computer coding scales, respectively. (That is, 26 = 64, 27 = 128, 28 = 256, 29= 512, and 210=1024.) In such numerical formats, the image data can lie readily analyzed with the aid of a computer. The use of computer assisted analysis techniques permits the spectral patterns in remote sensing data to be more fully examined. It also permits the data analysis process to be largely automated, providing cost advantages over visual interpretation techniques. However, just as humans are somewhat limited in their ability to interpret spectral patterns, computers are somewhat limited in their ability to evaluate spatial patterns. Therefore, visual and numerical techniques are complementary in nature, and consideration must be given to which approach (or combination of approaches) best fits a particular application. REFERENCE DATAAs we have indicated in the previous discussion, rarelyif everis remote sensing employed without the use of some form of reference data. The acquisition of reference data involves collecting measurements or observations about the objects, areas, or phenomena that are being sensed remotely. These data can take on any of a number of different forms and may be derived from a number of sources. For example, the data needed for a particular analysis might be derived from a soil survey map, a water quality laboratory report, or an aerial photograph. They may also stem from a field check on the identity, extent, and condition of agricultural crops, land uses, tree species, or water pollution problems. Reference data may also involve field measurements of temperature and other physical and/or chemical properties of various features. Reference data are often referred to by the term ground truth. This term is not meant literally, since many forms of reference data are not collected on the ground and can only approximate the truth of actual ground conditions. For example, ground truth may be Collected in the air, in the form of detailed aerial photographs used as reference data when analyzing less detailed high altitude or satellite imagery. Similarly, the ground truth will actually be water truth if we are studying water features. In spite of these inaccuracies, ground truth is a widely used term for reference data. Reference data might be used to serve any or all of the following purposes: 1. to aid in tile analysis and interpretation of remotely sensed data. 2. to calibrate a sensor. 3. to verify information extracted from remote sensing data. Hence, reference data must often be collected in accordance with the principles of statistical sampling design. Reference data can be very expensive and time consuming to collect properly. They can consist of either time-critical and/or time-stable measurements. Time-critical measurements are those made in ceases where ground conditions change rapidly with time, such as in the analysis of vegetation condition or water pollution events. Time- stable measurements are involved when the materials under observation do not change appreciably with time. For example, geologic applications often entail field observations that can be conducted at any time and that would not change appreciably from mission to mission. One form of reference data collection is the ground-based measurement of the reflectance and/or emittance of surface materials to determine their spectral response patterns. This might be done in the laboratory or in the field, using the principles of spectroscopy. Spectroscopic measurement procedures can involve the use of a variety of instruments. Often, a spectroradiometer is used in such measurement procedures. This device measures, as a function of wavelength, the energy coming from an object within its view. It is used primarily to prepare spectral reflectance curves lot various objects. In Laboratory spectroscopy, artificial sources of energy might be used to illuminate objects under study. In the Iab, other field parameters such as viewing geometry between object and sensor are also simulated. More often, therefore, in situ field measurements are preferred because of the many variables of the natural environment that influence remote sensor data that are difficult, if not impossible, to duplicate in the Laboratory. In the acquisition of field measurements, spectroradiometers may be operated in a number of modes, ranging from hand-held to helicopter or aircraft mounted. Figure 1.12 illustrates a highly portable instrument that is well suited to hand-held operation. This particular system acquires a continuous spectrum by recording data in 256 bands simultaneously. The spectral output goes to a microprocessor-based controller that records data on a built-in tape deck, and also displays and communicates the data in a standard computer-compatible format. Figure 1.13 shows a multiband radiometer that measures radiation in a series of discrete spectral bands, rather than over a continuous range. This particular device operates in eight spectral bands, seven of which match those used by the Thematic Mapper sensor onboard the Landsat satellites. The instrument is shown here suspended from a truck-mounted telescoping boom. Mounted in this manner, the radiometer can be driven to multiple field locations where spectral response measurements can be made quite conveniently. All data are again stored using a microprocessor-based data logger (located in the cab of the truck). Using a radiometer to obtain spectral reflectance measurements is normally a three-step process. First, the instrument is aimed at a calibration panel of known, stable reflectance. The purpose of this step is to quantify the incoming radiation or irradiance incident upon the measurement site. Next, the instrument is suspended over the target of interest and the radiation reflected by the object is measured. Finally, the spectral reflectance of the object is computed by ratioing the reflected energy measurement in each hand of observation to the incoming radiation measured in each band. Normally, the term reflectance factor is used to refer to the result of such computations. A reflectance factor is defined formally as the ratio of the radiant flux actually reflected by a sample surface to that which would be reflected into the same sensor geometry by an ideal, perfectly diffuse (Lamhertian) surface irradiated in exactly the same way as the sample.DIGITAL IMAGE PROCESSINGDigital image processing involves the manipulation and interpretation of digital images with the aid of a computer. This form of remote sensing actually began in the 1960s with a limited number of researchers analyzing airborne multi-spectral scanner data and digitized aerial photographs. However, it was not until the launch of Landsat-1, in 1972, that digital image data became widely available for land remote sensing applications. At that time, not only was the theory and practice of digital image processing in its infancy, the cost of digital computers was very high and their computational efficiency was very low by modern standards. Today, access to low cost, efficient computer hardware and software is commonplace and the sources of digital image data are many and varied. These sources range from commercial earth resource satellite systems, to the meteorological satellites, to airborne scanner data, to airborne solid-state camera data, to image data generated by scanning microdensitometers and high-resolution video cameras. All of these forms of data can be processed and analyzed using the techniques described in this chapter. Digital image processing is an extremely broad subject and it often involves procedures which can be mathematically complex. Our objective in this chapter is to introduce the basic principles of digital image processing without delving into the detailed mathematics involved. Also, we avoid extensive treatment of the hardware associated with digital image processing. The references at the end of this chapter are provided for those wishing to pursue such additional detail. The central idea behind digital image processing is quite simple. The digital image is fed into a computer one pixel at a time. The computer is programmed to insert these data into an equation, or series of equations, and then store the results of the computation for each pixel. These results form a new digital image that may be displayed or recorded in pictorial format or may itself be further manipulated by additional programs. The possible forms of digital image manipulation are literally infinite. However, virtually all these procedures may be categorized into one (or more) of the following four broad types of computer assisted operations:1. Image rectification and restoration. These operations aim to correct distorted or degraded image data to create a more faithful representation of the original scene. This typically involves the initial processing of raw image data to correct for geometric distortions, to calibrate tile data radiometrically, and to eliminate noise present in the data. Thus, the nature of any particular image restoration process is highly dependent upon the characteristics of the sensor used to acquire the image data. Image rectification and restoration procedures are often termed preprocessing operations because they normally precede further manipulation and analysis of the image data to extract specific information. We briefly discuss these procedures in Section 10.2 with treatment of various geometric corrections, radiometric corrections, and noise corrections.2. Image enhancement. These procedures are applied to image data in order to more effectively display or record the data for subsequent visual interpretation. Normally, image enhancement involves techniques for increasing the visual distinction between features in a scene. Tire objective is to create new images from the original image data in order to increase the amount of information that can be visually interpreted from the data. The enhanced images can be displayed interactively on a monitor or they can be recorded in a hard copy format, either in black and white or in color. There are no simple rules for producing the single best image for a particular application. Often several enhancements made from the same raw image are necessary. We summarize the various broad approaches to enhancement in Section 10.3. In Section 10.4, we treat specific procedures that manipulate the contrast of an image (level slicing and contrast stretching). In Section 10.5, we discuss spatial feature manipulation (spatial filtering, convolution, edge enhancement, and Fourier analysis). In Section 10.6, we consider enhancements involving multiple spectral bands of imagery (spectral ratioing, principal and canonical components, vegetation components, and intensity-hue-saturation color space transformations). 3. Image classification. The objective of these operations is to replace visual analysis of the image data with quantitative techniques for automating the identification of features in a scene. This normally involves the analysis of multispectral image data and the application of statistically based decision rules for det
展开阅读全文