This article discusses the theoretical and practical aspects of using hyperspectral imaging data for automatic target detection. It focuses on techniques that use spectral information to classify each pixel as either a target or non-target. The article explains the structure of hyperspectral data, the development of detection algorithms, and the performance metrics used to evaluate these algorithms. It also presents a taxonomy of detection algorithms and provides results for their theoretical performance. The article uses data from the HYDICE and Hyperion sensors to illustrate the operation and performance of various detectors. Hyperspectral imaging sensors measure the spectral radiance of materials in each pixel, providing both spatial and spectral information. The reflectance spectrum, which is the ratio of reflected to incident radiation, varies with wavelength. The presence of the atmosphere and the nature of the solar spectrum affect the relationship between the observed radiance and the reflectance spectra. The measured radiance spectrum is the solar spectrum modified by atmospheric transmittance and the reflectance of the imaged surface. The article discusses the challenges of measuring the spectral properties of materials through the atmosphere, including atmospheric absorption and scattering, illumination effects, and sensor response. It explains how sophisticated atmospheric compensation codes can be used to recover the reflectance spectrum from the observed radiance spectrum. The resulting spectral signature can be used to identify specific materials in a scene. The article also discusses the differences between spatial and spectral processing in remote sensing. It explains that hyperspectral sensors have higher spectral resolution than multispectral sensors, allowing for more detailed spectral information. The article discusses the challenges of spectral variability and mixed-pixel interference in hyperspectral data, which can affect the accuracy of target detection. The article presents three families of mathematical models used to characterize spectral variability in hyperspectral data: probability density models, subspace models, and linear spectral mixing models. These models are used to describe the variability of spectra and to develop detection algorithms. The article discusses the design, evaluation, and taxonomy of target detectors. It explains that detection algorithms are based on statistical models and that the performance of these algorithms is evaluated using metrics such as the probability of detection and the probability of false alarm. The article also discusses the challenges of designing detectors for practical applications, including the need for adaptive detectors and the use of constant false-alarm-rate (CFAR) processors. The article concludes by discussing the challenges of detecting targets in hyperspectral data, including the need for accurate statistical models and the impact of spectral variability and mixed-pixel interference on detection performance. It emphasizes the importance of developing robust and efficient detection algorithms for hyperspectral imaging applications.This article discusses the theoretical and practical aspects of using hyperspectral imaging data for automatic target detection. It focuses on techniques that use spectral information to classify each pixel as either a target or non-target. The article explains the structure of hyperspectral data, the development of detection algorithms, and the performance metrics used to evaluate these algorithms. It also presents a taxonomy of detection algorithms and provides results for their theoretical performance. The article uses data from the HYDICE and Hyperion sensors to illustrate the operation and performance of various detectors. Hyperspectral imaging sensors measure the spectral radiance of materials in each pixel, providing both spatial and spectral information. The reflectance spectrum, which is the ratio of reflected to incident radiation, varies with wavelength. The presence of the atmosphere and the nature of the solar spectrum affect the relationship between the observed radiance and the reflectance spectra. The measured radiance spectrum is the solar spectrum modified by atmospheric transmittance and the reflectance of the imaged surface. The article discusses the challenges of measuring the spectral properties of materials through the atmosphere, including atmospheric absorption and scattering, illumination effects, and sensor response. It explains how sophisticated atmospheric compensation codes can be used to recover the reflectance spectrum from the observed radiance spectrum. The resulting spectral signature can be used to identify specific materials in a scene. The article also discusses the differences between spatial and spectral processing in remote sensing. It explains that hyperspectral sensors have higher spectral resolution than multispectral sensors, allowing for more detailed spectral information. The article discusses the challenges of spectral variability and mixed-pixel interference in hyperspectral data, which can affect the accuracy of target detection. The article presents three families of mathematical models used to characterize spectral variability in hyperspectral data: probability density models, subspace models, and linear spectral mixing models. These models are used to describe the variability of spectra and to develop detection algorithms. The article discusses the design, evaluation, and taxonomy of target detectors. It explains that detection algorithms are based on statistical models and that the performance of these algorithms is evaluated using metrics such as the probability of detection and the probability of false alarm. The article also discusses the challenges of designing detectors for practical applications, including the need for adaptive detectors and the use of constant false-alarm-rate (CFAR) processors. The article concludes by discussing the challenges of detecting targets in hyperspectral data, including the need for accurate statistical models and the impact of spectral variability and mixed-pixel interference on detection performance. It emphasizes the importance of developing robust and efficient detection algorithms for hyperspectral imaging applications.