To increase productivity in agricultural production, speed, and accuracy is the key requirement for long-term economic growth, competitiveness, and sustainability. Traditional manual paddy rice seed classification operations are costly and unreliable because human decisions in identifying objects and issues are inconsistent, subjective, and slow. Machine vision technology provides an alternative for automated processes, which are nondestructive, cost-effective, fast, and accurate techniques. In this work, we presented a study that utilized machine vision technology to classify 14 Oryza sativa rice varieties. Each cultivar used over 3,500 seed samples, a total of close to 50,000 seeds. There were three main processes, including preprocessing, feature extraction, and rice variety classification. We started the first process using a seed orientation method that aligned the seed bodies in the same direction. Next, a quality screening method was applied to detect unusual physical seed samples. Their physical information including shape, color, and texture properties was extracted to be data representations for the classification. Four methods (LR, LDA, k-NN, and SVM) of statistical machine learning techniques and five pretrained models (VGG16, VGG19, Xception, InceptionV3, and InceptionResNetV2) on deep learning techniques were applied for the classification performance comparison. In our study, the rice dataset were classified in both subgroups and collective groups for studying ambiguous relationships among them. The best accuracy was obtained from the SVM method at 90.61%, 82.71%, and 83.9% in subgroups 1 and 2 and the collective group, respectively, while the best accuracy on the deep learning techniques was at 95.15% from InceptionResNetV2 models. In addition, we showed an improvement in the overall performance of the system in terms of data qualities involving seed orientation and quality screening. Our study demonstrated a practical design of rice classification using machine vision technology.

No abstract

The optical spectrometer is a widely established and widely used scientific instrument for analyzing biochemical samples. However, it is expensive, bulky, and requires an external power source and a computer to operate, which makes it difficult to use in resourcelimited countries and in remote locations. Advancement in smartphone technology has led to the development of portable, low-cost smartphone-based spectrometers. However, the lack of smartphonebased software and automatic wavelength calibration results in irrelevant workloads for students within the time limit of the course. In addition, most smartphone spectrometers are only able to capture images. As a result, a computer becomes an essential tool for the student to process and analyze the sample spectrum. Unlike many other smartphone-based spectrometers, our developed device also provides a smartphone application for both Android and iOS platforms. Thus, measurement data and data plots are transferrable among smartphones and computers. In addition, the wavelength calibration is carried out automatically using an embedded reference white light LED by the application. Furthermore, the provided 3D-printed parts, part lists, and circuit board are included with an example of an iPhone 7 mount, as well as their assembly instructions. Our developed device features a sufficiently quantitative measurement for wavelength ranges from 411.2 to 700 nm. For the spectral resolution, our device achieved resolutions of 9.8, 13.4, and 21.7 nm at wavelengths of 405, 532, and 650 nm, respectively, when evaluated with an iPhone 7.

In this paper, a novel and simple method based on an all-pass filter is presented by using the benefits of a pole re-position technique. The gains at frequencies of π of each single notch filter are adjusted by adding N − 1 tuning variables. The proposed method ensures that the passband gains are uniformly flat, the notch frequencies exactly meet the specifications, and the realized 3 dB bandwidths are approximately the same as those specified. This technique is very useful for designers because there are only N − 1 tuning variables required to adjust, and it does not need a complicated mathematical calculation. Although this technique is simple and easy to implement, it requires many iterations to find a suitable parameter for making the passband gain between two notch frequencies uniformly flat. Hence in this paper, three searching algorithms have been applied to reduce the number of iterations. We obtain a faster search and a closer frequency response to the ideal one.

As fast human temperature screening is needed in large public areas, this paper proposes a low-cost mobile platform module that combines the advantages of analyzing visible and thermal images. In particular, the key idea relies on face detection in the visible image. Then the coordinates of all faces detected are mapped on to the thermal image to determine their corresponding temperatures. Internal temperature compensation and external reference temperature also are employed to reduce the unwanted temperature fluctuation inside the module and in the surrounding environment. Our mobile platform module, called , uses a FLIR ONE camera as our visible and thermal imaging cameras. It can simultaneously determine the temperatures of nine people at a speed of 8 frames/second. A field test operation was performed for four days with 1,170 people, with very promising results of 100% sensitivity, 92.6% specificity, and 92.7% accuracy.