Rice leaf infections are a common hazard to rice production, affecting many farmers all over the world. Early detection and treatment of rice leaf infection are critical for promoting healthy rice plant growth and ensuring adequate supply for the fast-growing population. Computer-assisted rice leaf disease diagnoses are hampered due to strong image backgrounds. Popular Convolutional Neural Network (CNN) architecture extracts the features from images and diagnoses the disease to address the issues above. However, this method is best suitable for segmented images and gives low accuracy with real-time images. In this case, the Internet of Things is a paradigm shift that collects agro-meteorological information that effectively helps diagnose rice diseases. Motivated by the usefulness of CNN models and agricultural IoT, a novel multimodal data fusion framework named Rice-Fusion is proposed to diagnose rice disease. Rice disease diagnosis based on a single modality may not be accurate, and hence the fusion of heterogeneous modalities is essential for robust and reliable disease diagnosis. This gives a new dimension to the domain of rice disease diagnosis. The dataset was collected manually with 3200 rice health category samples using two modalities, namely agro-meteorological sensors and a camera. The Rice-Fusion framework initially extracts the numerical features from agro-meteorological data collected from sensors. Next, it extracts the visual features from the captured rice images. These extracted features are further fused using a concatenation layer followed by a dense layer, which provides single output for diagnosing the rice disease. The testing accuracy of Rice-Fusion is 95.31% as opposed to other unimodal framework accuracies of 82.03% and 91.25% based on CNN and Multi-Layer Perceptron (MLP) architectures, respectively. Experimental results analysis demonstrates that the proposed Rice-Fusion multimodal data fusion framework outperforms the outcome of unimodal frameworks.

https://ieeexplore.ieee.org/ielx7/6287639/9668973/09672157.pdf

Rice-Fusion: A Multimodality Data Fusion Framework for Rice Disease Diagnosis

Rice leaf infections are a common hazard to rice production, affecting many farmers all over the world. Early detection and treatment of rice leaf infection are critical for promoting healthy rice plant growth and ensuring adequate supply for the fast-growing population. Computer-assisted rice leaf disease diagnoses are hampered due to strong image backgrounds. Popular Convolutional Neural Network (CNN) architecture extracts the features from images and diagnoses the disease to address the issues above. However, this method is best suitable for segmented images and gives low accuracy with real-time images. In this case, the Internet of Things is a paradigm shift that collects agro-meteorological information that effectively helps diagnose rice diseases. Motivated by the usefulness of CNN models and agricultural IoT, a novel multimodal data fusion framework named Rice-Fusion is proposed to diagnose rice disease. Rice disease diagnosis based on a single modality may not be accurate, and hence the fusion of heterogeneous modalities is essential for robust and reliable disease diagnosis. This gives a new dimension to the domain of rice disease diagnosis. The dataset was collected manually with 3200 rice health category samples using two modalities, namely agro-meteorological sensors and a camera. The Rice-Fusion framework initially extracts the numerical features from agro-meteorological data collected from sensors. Next, it extracts the visual features from the captured rice images. These extracted features are further fused using a concatenation layer followed by a dense layer, which provides single output for diagnosing the rice disease. The testing accuracy of Rice-Fusion is 95.31&#x0025; as opposed to other unimodal framework accuracies of 82.03&#x0025; and 91.25&#x0025; based on CNN and Multi-Layer Perceptron (MLP) architectures, respectively. Experimental results analysis demonstrates that the proposed Rice-Fusion multimodal data fusion framework outperforms the outcome of unimodal frameworks. 

We have developed a completely new type of general-purpose CCD data acquisition system which enables one to drive any type of CCD using any type of clocking mode. A CCD driver system widely used before consisted of an analog multiplexer (MPX), a digital-to-analog converter (DAC), and an operational amplifier. A DAC is used to determine high and low voltage levels and the MPX selects each voltage level using a TTL clock. In this kind of driver board, it is difficult to reduce the noise caused by a short of high and low level in MPX and also to select many kinds of different voltage levels. Recent developments in semiconductor IC enable us to use a very fast sampling ($\sim$ 10MHz) DAC with low cost. We thus develop the new driver system using a fast DAC in order to determine both the voltage level of the clock and the clocking timing. We use FPGA (Field Programmable Gate Array) to control the DAC. We have constructed the data acquisition system and found that the CCD functions well with our new system. The energy resolution of Mn K$\alpha$ has a full-width at half-maximum of $\simeq$ 150 eV and the readout noise of our system is $\simeq$ 8 e$^-$.

Fast and Flexible CCD Driver System Using Fast DAC and FPGA

Dynamic optical imaging (e.g. time delay integration imaging) is troubled by the motion blur fundamentally arising from mismatching between photo-induced charge transfer and optical image movements. Motion aberrations from the forward dynamic imaging link impede the acquiring of high-quality images. Here, we propose a high-resolution dynamic inversion imaging method based on optical flow neural learning networks. Optical flow is reconstructed via a multilayer neural learning network. The optical flow is able to construct the motion spread function that enables computational reconstruction of captured images with a single digital filter. This works construct the complete dynamic imaging link, involving the backward and forward imaging link, and demonstrates the capability of the back-ward imaging by reducing motion aberrations.

/pdf/high-resolution-dynamic-inversion-imaging-with-motion-25c3806rkx.pdf

High-resolution dynamic inversion imaging with motion-aberrations-free using optical flow learning networks.

The existence of blemishes deteriorates the imaging quality of fiber-optic imaging elements and reduces the yield of finished products in factories. The effective detection of blemishes is a prerequisite for analyzing the causes of blemishes and preventing their generation. An independently developed detection device for fiber-optic imaging elements based on machine vision is implemented to detect blemishes automatically and accurately. The blemish distributions of three typical fiber-optic imaging elements, including a fiber-optic plate (FOP), a fiber-optic inverter (FOI), and a fiber-optic taper (FOT), are compared with each other, and the causes of blemishes in fiber-optic imaging elements are analyzed. The distribution of blemishes in the FOP is random, and the average number (AN) of blemishes in the first zone (central zone) and the second zone (outer zone) with the same area is similar. The AN of blemishes for the FOI prepared by twisting the FOP at high-temperature increases from 272.8 (FOP) to 2125.0, and the AN in the outer zone is 1542.5, higher than that in the central zone. More blemishes are generated by the twisting process, and the distribution of these blemishes is close to the outer zone. A significant amount of chicken wires is found in the outer zone, but the majority of blemishes are still spots. The AN of blemishes for the FOT prepared by stretching the FOP increases from 383.0 (FOP) to 515.0, which is attributed to the increase of the number of blemishes in the first zone from 190.8 to 427.0, but the AN of blemishes in the second zone decreases from 192.3 to 88.0. The stretching process stimulates the formation of blemishes with a distribution that is close to the central ring. These blemishes are randomly distributed inside or at the boundary of the multifiber. Most of the blemishes are still spots.

Detection of blemish for fiber-optic imaging elements

In this paper, the performance of the electronic conventional image motion compensation (IMC) method based on the time delay integration (TDI) mode was analyzed using the optical injection formula of charge coupled devices (CCDs). The result shows that the non-synchronous effect of charge packet transfer caused by line-by-line transfer during exposure makes the compensated image dissatisfying. Then an improved electronic IMC method based on the CCD multiphase structure was proposed. In this method, a series of proper driving clocks were applied to drive the charge packet to move electrode-by-electrode during the exposure time, which results in a minimum non-synchronous effect of charge packet transfer. The mismatch of velocity between charge packet transfer and image motion was decreased. The performance of the improved electronic IMC method was also analyzed using the optical injection formula. The modulation degrees of the two methods were compared. The average value of the modulation degree of the improved electronic IMC method was 47/96, greater than the conventional electronic IMC method, which was 1/3. To achieve the improved electronic IMC, the driver timing diagram of the improved electronic IMC method was proposed. This paper presented an improved hardware implementation method for the improved electronic IMC method. Based on the basic FTF4052M drive circuit system, an IMC pulse pattern generator that worked together with the main pulse pattern generator (SAA8103) was added to achieve the improved electronic IMC. Then, the internal structure of the IMC pulse pattern generator was given. A dual pulse pattern generator drive circuit system was proposed. After computer simulation and indoor real shot verification, the compensation effect of the improved electronic IMC method was better than the compensation effect of the conventional electronic IMC method.

An Improved Electronic Image Motion Compensation (IMC) Method of Aerial Full-Frame-Type Area Array CCD Camera Based on the CCD Multiphase Structure and Hardware Implementation.

This paper addresses the lack of robustness of feature selection algorithms for fuzzy clustering segmentation with the Gaussian mixture model Assuming that the neighbourhood pixels and the centre pixels obey the same distribution, a Markov method is introduced to construct the prior probability distribution and achieve the membership degree regularisation constraint for clustering sample points Then, a noise smoothing factor is introduced to optimise the prior probability constraint Second, a power index is constructed by combining the classification membership degree and prior probability since the Kullback–Leibler (KL) divergence of the noise smoothing factor is used to supervise the prior probability; this probability is embedded into Fuzzy Superpixels Fuzzy C-means (FSFCM) as a regular factor This paper proposes a fuzzy clustering image segmentation algorithm based on an adaptive feature selection Gaussian mixture model with neighbourhood information constraints To verify the segmentation performance and anti-noise robustness of the improved algorithm, the fuzzy C-means clustering algorithm Fuzzy C-means (FCM), FSFCM, Spatially Variant Finite Mixture Model (SVFMM), EGFMM, extended Gaussian mixture model (EGMM), adaptive feature selection robust fuzzy clustering segmentation algorithm (AFSFCM), fast and robust spatially constrained Gaussian mixture model (GMM) for image segmentation (FRSCGMM), and improve method are used to segment grey images containing Gaussian noise, salt-and-pepper noise, multiplicative noise and mixed noise The peak signal-to-noise ratio (PSNR) and the error rate (MCR) are used as the theoretical basis for assessing the segmentation results The improved algorithm indicators proposed in this paper are optimised The improved algorithm yields increases of 01272–129803 dB, 15501–134396 dB, 19113–112613 dB and 10233–102804 dB over the other methods, and the Misclassification rate (MSR) decreases by 032–3732%, 502–4105%, 03–2179% and 09–3095% compared to that with the other algorithms It is verified that the segmentation results of the improved algorithm have good regional consistency and strong anti-noise robustness, and they meet the needs of noisy image segmentation

An Adaptive Feature Selection Algorithm for Fuzzy Clustering Image Segmentation Based on Embedded Neighbourhood Information Constraints.

Since the fuzzy local information C-means (FLICM) segmentation algorithm cannot take into account the impact of different features on clustering segmentation results, a local fuzzy clustering segmentation algorithm based on a feature selection Gaussian mixture model was proposed. First, the constraints of the membership degree on the spatial distance were added to the local information function. Second, the feature saliency was introduced into the objective function. By using the Lagrange multiplier method, the optimal expression of the objective function was solved. Neighborhood weighting information was added to the iteration expression of the classification membership degree to obtain a local feature selection based on feature selection. Each of the improved FLICM algorithm, the fuzzy C-means with spatial constraints (FCM_S) algorithm, and the original FLICM algorithm were then used to cluster and segment the interference images of Gaussian noise, salt-and-pepper noise, multiplicative noise, and mixed noise. The performances of the peak signal-to-noise ratio and error rate of the segmentation results were compared with each other. At the same time, the iteration time and number of iterations used to converge the objective function of the algorithm were compared. In summary, the improved algorithm significantly improved the ability of image noise suppression under strong noise interference, improved the efficiency of operation, facilitated remote sensing image capture under strong noise interference, and promoted the development of a robust anti-noise fuzzy clustering algorithm.

A Local Neighborhood Robust Fuzzy Clustering Image Segmentation Algorithm Based on an Adaptive Feature Selection Gaussian Mixture Model

When using a high-resolution full-frame area array charge-coupled device (CCD) FTF4052M as the aerial image sensor of a camera, the frame rate is generally below 1  frame  /  second (fps), which cannot meet the requirements of applications with high frame rates. First, the FTF4052M drive circuit is introduced. Next, an improved driver circuit of the CCD 4052M is proposed. Using the CCD to output four amplifiers simultaneously yielded a maximum frame rate of 3.4 fps. Then, the timing of the CCD driver, the front-end processing circuit, the DC bias circuit, and the interface circuit of the four outputs are designed. The improved driver circuit could meet the application requirements of various aerial cameras. In addition, the FTF4052M nonuniformity is analyzed, and a nonuniformity response detection system is established. Using this system, the FTF4052M array’s nonuniformity among the four quadrants and among the pixels is tested. Based on the CCD’s linear responsivity, two correction algorithms are proposed to correct the nonuniformity. Finally, using the correction, the four quadrants’ standard deviations of the response sensitivity are reduced by a factor of 1  /  13, and the responsive nonuniformity among all pixels is reduced by a factor of 1  /  10. Through the reshooting of the identified rate board, it is found that the nonuniformity of the CCD array is markedly improved.

Design of a high frame rate driver circuit for a 22 M-pixel high-resolution large-area array CCD camera and its nonuniformity correction

In order to design a charge-coupled device (CCD) driving circuit with a high frame rate and a large area array which can support the electronic image motion compensation function, first a basic dri...

Hang Ren

Papers

An Improved Electronic Image Motion Compensation (IMC) Method of Aerial Full-Frame-Type Area Array CCD Camera Based on the CCD Multiphase Structure and Hardware Implementation.

An Adaptive Feature Selection Algorithm for Fuzzy Clustering Image Segmentation Based on Embedded Neighbourhood Information Constraints.

A Local Neighborhood Robust Fuzzy Clustering Image Segmentation Algorithm Based on an Adaptive Feature Selection Gaussian Mixture Model

Design of a high frame rate driver circuit for a 22 M-pixel high-resolution large-area array CCD camera and its nonuniformity correction

Driving circuitry of a full-frame area array charge-coupled device (CCD) supporting multiple output modes and electronic image motion compensation