Automatic target detection process
First Claim
Patent Images
1. A method for processing imaging data gathered by an imaging sensor to detect the presence or absence of a target, the imaging sensor including pulsed light source means and gated camera means, the method comprising the steps of:
- converting imaging data received by said gated camera means into an array of pixels "n";
determining physical parameters related to environmental conditions and imaging sensor location and orientation;
inputting said physical parameters into a hypothetical physical and optical model and calculating the hypothetical distribution of the number of photons impinging on said gated camera means for each of said pixels "n" under each of two hypotheses including a first hypothesis where no target is present and a second hypothesis where a target is present;
assuming that the impinging photons are gaussian in both the first and second hypotheses, calculating the mean μ
n 0 and variance σ
n 0 in each pixel "n" when no target is present and calculating the mean μ
n,mA and variance σ
n,m4 in each pixel "n" when a target is in the state of hypothesis based on said hypothetical distribution of photons for each pixel;
digitizing said pixels to define a digitized data input;
defining a set of sub hypotheses within said second hypothesis, wherein said sub hypotheses represent possible target states,calculating likelihood ratios for each of the sub hypotheses with respect to the digitized data input;
multiplying the likelihood ratio of each hypothesis by a prior probability of the hypothesis to obtain intermediate probabilities;
multiplying each intermediate probability by a corresponding element in a Bayesian gain matrix and summing the resultant products to estimate the gain associated with each of said sub hypotheses; and
applying an optimal statistical decision rule to the gain estimated for each of the sub hypotheses and comparing the resultant data to a preselected threshold value to determine the presence or absence of a target.
1 Assignment
0 Petitions
Accused Products
Abstract
A novel data processing technique is provided for detecting, locating and identifying targets from a plurality of images generated by an imaging sensor such as an imaging lidar system. The present invention employs physical models of signals produced by target objects of interest. Such a model based detection system globally processes frames of data to determine the existence and location of component elements that characterize the target being modeled.
108 Citations
29 Claims
-
1. A method for processing imaging data gathered by an imaging sensor to detect the presence or absence of a target, the imaging sensor including pulsed light source means and gated camera means, the method comprising the steps of:
-
converting imaging data received by said gated camera means into an array of pixels "n"; determining physical parameters related to environmental conditions and imaging sensor location and orientation; inputting said physical parameters into a hypothetical physical and optical model and calculating the hypothetical distribution of the number of photons impinging on said gated camera means for each of said pixels "n" under each of two hypotheses including a first hypothesis where no target is present and a second hypothesis where a target is present; assuming that the impinging photons are gaussian in both the first and second hypotheses, calculating the mean μ
n 0 and variance σ
n 0 in each pixel "n" when no target is present and calculating the mean μ
n,mA and variance σ
n,m4 in each pixel "n" when a target is in the state of hypothesis based on said hypothetical distribution of photons for each pixel;digitizing said pixels to define a digitized data input; defining a set of sub hypotheses within said second hypothesis, wherein said sub hypotheses represent possible target states, calculating likelihood ratios for each of the sub hypotheses with respect to the digitized data input; multiplying the likelihood ratio of each hypothesis by a prior probability of the hypothesis to obtain intermediate probabilities; multiplying each intermediate probability by a corresponding element in a Bayesian gain matrix and summing the resultant products to estimate the gain associated with each of said sub hypotheses; and applying an optimal statistical decision rule to the gain estimated for each of the sub hypotheses and comparing the resultant data to a preselected threshold value to determine the presence or absence of a target. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23)
-
-
24. A method for processing imaging data gathered by an imaging sensor to detect the presence of absence of a target, the imaging sensor including pulsed light source means and gated camera means, the method comprising the steps of:
-
converting imaging data received by said gated camera means into an array of pixels "n"; determining physical parameters related to environmental conditions and imaging sensor location and orientation; inputting said physical parameters into a hypothetical physical and optical model and calculating the hypothetical distribution of the number of photons impinging on said gated camera means for each of said pixels "n" under each of two hypotheses including a first hypothesis where no target is present and a second hypothesis where a target is present; assuming that the impinging photons are gaussian in both the first and second hypotheses, calculating the mean μ
n 0 and variance σ
n 0 in each pixel "n" where no target is present and calculating the mean μ
n,mA and variance σ
n,mA in each pixel "n" when a target is in the state of hypothesis based on said hypothetical distribution of photons for each pixel;digitizing said pixels to define a digitized data input; calculating likelihood ratios for each of the hypotheses with respect to the digitized data input; multiplying the likelihood ratio of each hypothesis by a prior probability of the hypothesis and determining the maximum of the products; and comparing the maximum to a preselected threshold value to determine the presence of absence of a target.
-
-
25. A method for processing imaging data gathered by an imaging sensor to detect the presence or absence of a target, the imaging sensor including pulsed light source means and gated camera means, the method comprising the steps of:
-
converting imaging data received by said gated camera means into an array of pixels "n"; deriving a set of feasible hypothesis for the state of a possible target under each of two hypotheses including a first hypothesis where no target is present and a second hypothesis where a target is present; calculating themes μ
n 0 and variance σ
n 0 in each of said pixels "n" when no target is present and calculating the mean μ
n,mA and variance σ
n,mA in each of said pixels "n" when a target is in the state of hypothesis;digitizing said pixels to define a digitized data input; defining a set of sub hypotheses within said second hypothesis, wherein said sub hypotheses represent possible target states; calculating likelihood ratios for each of the sub hypotheses with respect to the digitized data input; multiplying the likelihood ratio of each hypothesis by a prior probability of the hypothesis to obtain intermediate probabilities; multiplying each intermediate probability by a corresponding element in a Bayesian gain matrix and summing the resultant products to estimate the gain associated with each of said sub hypotheses; and applying an optimal statistical decision rule to the gain estimated for each of the sub hypotheses and comparing the resultant data to a preselected threshold value to determine the presence or absence of a target. - View Dependent Claims (26)
-
-
27. A method for processing imaging data gathered by an imaging sensor to detect the presence or absence of a target, the imaging sensor including pulsed light source means and gated camera means, the method comprising the steps of:
-
converting imaging data received by said gated camera means into an array of pixels "n"; deriving a set of feasible hypotheses for the state of a possible target under each of two hypotheses including a first hypothesis where no target is present and a second hypothesis where a target is present; assuming that the impinging photons are gaussian in both the first and second hypotheses, calculating the mean μ
n 0 and variance σ
n 0 in each of said pixels "n" where no target is present and calculating the mean μ
n,mA and variance σ
n,mA in each of said pixels "n" when a target is in the state of hypothesis;digitizing said pixels to define a digitized data input; calculating likelihood ratios for each of the hypotheses with respect to the digitized data input; multiplying the likelihood ratio of each hypothesis by a prior probability of the hypothesis and determining the maximum of the products; and comparing the maximum to a preselected threshold value to determine the presence of absence of target.
-
-
28. An apparatus for processing imaging data gathered by an imaging sensor to detect the presence of absence of a target, the imaging sensor including pulsed light source means and gated camera means, comprising:
-
means for converting imaging data received by said gated camera means into an array of pixels "n"; means for determining physical parameters related to environmental conditions and imaging sensor location and orientation; means for inputting said physical parameters into a hypothetical physical and optical model and calculating the hypothetical distribution of the number of photons impinging on said gated camera means for each of said pixels "n" under each of two hypotheses including a first hypothesis where no target is present and a second hypothesis where a target is present; means for calculating the mean μ
n 0 and variance σ
n 0 in each pixel "n" when no target is present and calculating the mean μ
n,mA and variance σ
n,mA in each pixel "n" when a target is in the state of hypothesis based on said hypothetical distribution of photons for each pixel;means for digitizing said pixels to define a digitized data input; means for defining a set of sub hypotheses within said second hypothesis, wherein said sub hypotheses represent possible target states; means for calculating likelihood ratios for each of the hub hypotheses with respect to the digitized data input; means for multiplying the likelihood ratio of each hypothesis by a prior probability of the hypothesis to obtain intermediate probabilities; means for multiplying each intermediate probability by a corresponding element in a Bayesian gain matrix and summing the resultant products to estimate the gain associated with each of said sub hypotheses; and means for applying an optimal statistical decision rule to the gain estimated for each of the sub hypotheses and comparing the resultant data to a preselected threshold value to determine the presence or absence of a target.
-
-
29. A method for processing imaging data gathered by an imaging sensor to detect the presence or absence of a target, the imaging sensor including pulsed light source means and gated camera means, the method comprising:
-
converting imaging data received by said gated camera means into an array of pixels "n"; determining physical parameters related to environmental conditions and imaging sensor location and orientation; inputting said physical parameters into a hypothetical physical and optical model and calculating the hypothetical distribution of the number of photons impinging on said gated camera means for each of said pixels "n" under each of two hypotheses including a first hypothesis where no target is present and a second hypothesis where a target is present; calculating the mean μ
n 0 and variance σ
n 0 in each pixel "n" when no target is present and calculating the mean μ
n,mA and variance σ
n,mA in each pixel "n" when a target is in the state of hypothesis based on said hypothetical distribution of photons for each pixel;digitizing said pixels to define a digitized data input; defining a set of sub hypotheses within said second hypothesis, wherein said sub hypotheses represent possible target states; calculating likelihood ratios for each of the sub hypotheses with respect to the digitized data input; multiplying the likelihood ratio of each hypothesis by a prior probability of the hypothesis to obtain intermediate probabilities; multiplying each intermediate probability by a corresponding element in a Bayesian gain matrix and summing the resultant products to estimate the gain associated with each of said sub hypothesis; and applying an optimal statistical decision rule to the gain estimated for each of the sub hypotheses and comparing the resultant data to a preselected threshold value to determine the presence or absence of a target.
-
Specification