This example uses a data sample from the Jasper Ridge dataset as the test data. The test data contains four endmembers latent, consisting of roads, soil, water, and trees. In this example, you will:. Classify hyperspectral images using a custom spectral convolution neural network CSCNN for classification.
Identify the classes of endmember materials present in a hyperspectral image. The endmembers are pure spectral signature that signifies the reflectance characteristics of pixels belonging to a single surface material. The existing endmember extraction or identification algorithms extracts or identifies the pure pixels in a hyperspectral image. However, these techniques do not identify the material name or class to which the endmember spectrum belong to.
In this example, you will extract the endmember signatures and then, classify or identify the class of an endmember material in the hyperspectral image by using spectral matching. Detect a known target in the hyperspectral image by using spectral matching method. The pure spectral signature of the known target material is used to detect and locate the target in a hyperspectral image.
In this example, you will use the spectral angle mapper SAM spectral matching method to detect man-made roofing materials known target in a hyperspectral image. The spectral signatures of all the pixels in the data cube are compared with the reference spectrum and the best matching pixel spectrum is classified as belonging to the target material.
Identify the types of vegetations regions in a hyperspectral image through interactive thresholding of a normalized difference vegetation index NDVI map. The number of segmented regions was quantized using entropy. Therefore, entropy can represent the complexity of regions. Based on entropy, the optimal bands are selected.
The final detection results were obtained using -means clustering with the selected bands. Figure 2 b shows sample spectral band images. A camouflaged object detection problem can be regarded as selecting suitable spectral bands that discriminate interesting region in normal background. The proposed statistical distance metric can be useful to generate candidate bands because a hypercube image provides enough samples about millions of spectral profiles and statistical distance can measure the distinctiveness of spectral bands, which leads to easy detection of camouflaged objects.
For example, if a test hypercube consists of a real leaf and a printed leaf, the spectral profile and specific band image are obtained as shown in Figure 3. Statistical distance-based, candidate band selection is motivated by the observation of band image analysis, as shown in Figure 4 a. According to the distribution, two Gaussian distribution functions foreground and background parameterized from the means and standard deviations can be fitted.
The class discriminability measure is defined as. By applying the aforementioned equation to each band, the band discriminability curve can be obtained according to the wavelength, as shown in Figure 4 b. The candidate bands can be selected by applying local maxima or global maxima to the curve. If -means clustering with th band image is performed, the discriminability value can be obtained as mentioned above. At the same time, a segmented image using the class labels in image space can be acquired.
If a hypercube image has the size of samples scan length bands , the complexity of segmented regions at the th band can be quantified using entropy.
Entropy can measure the complexity of spatial region distribution. In the camouflaged object detection problem, the ideal number of regions is just two foreground and background. Therefore, high entropy can represent large number of segmented regions. The region entropy is defined as where is the probability of the pixels belonging to th region. This is defined as. Ideally, the detection results consist of one abnormal region and the other background region. If the number of segmented region increases, the region entropy increases.
Therefore, a threshold is applied for the region entropy to reduce the candidate bands that generate many small regions. Figure 5 shows the region segmentation results according to the different region entropy values. The region entropy threshold around 1 is normally used.
The proposed method was validated in terms of the band selection scheme using the same -means clustering unsupervised classifier. The baseline band selection method was principle components analysis PCA , which is an effective data reduction technique that is used frequently in hyperspectral data analysis [ 6 ].
In PCA, a human manually selects a principle component i. As a second baseline method, the entire spectrum curve, where all the bands are selected, is used [ 13 ]. The detection rate DR , false alarm rate FAR , and the number of bands used for quantitative comparison are used. Table 2 lists the overall performance comparison of the leaf database. PCA method selected 7 bands The Proposed 1 method selected 9 bands The Proposed 2 method with 4 selected bands Figures 7 c — 7 f show the qualitative performance comparison results for a given test hypercube Figure 7 a and a ground truth image Figure 7 b.
The Proposed 2 method could detect the camouflaged region perfectly. In terms of detection time complexity, the Proposed 2 method took only 0. The space complexity is proportional to the number of bands. So, the Proposed 2 method occupies the smallest memory space. Another test was conducted to validate the proposed method for the hair database, which consists of a wig and hair.
Walker , Patrick H. Worley , John B. Abstract The spectral transform method is a widely-used numerical technique for solving partial differential equations on the sphere in global climate modeling. Keyphrases spectral method reasonable amount computational requirement nonlinear shallow water equation current model partial differential equation global climate modeling climate model simulation parallel multiprocessor greenhouse gas widely-used numerical technique climate model spectral transform method node intel ipsc complex climate model computational kernel research program atmospheric concentration shallow water equation.
Has PDF. Publication Type. More Filters. Algebraic Methods for Log-Linear Models. Techniques from representation theory Diaconis, and algebraic geometry Drton et al. With these … Expand. View 1 excerpt. Harmonic analysis of homogeneous networks. IEEE Trans. Neural Networks.
0コメント