Detection of citrus diseases in complex backgrounds based on image-text multimodal fusion and knowledge assistance.

Diseases pose a significant threat to the citrus industry, and the accurate detection of these diseases represent key factors for their early diagnosis and precise control. Existing diagnostic methods primarily rely on image models trained on vast datasets and limited their applicability due to singular backgrounds. To devise a more accurate, robust, and versatile model for citrus disease classification, this study focused on data diversity, knowledge assistance, and modal fusion. Leaves from healthy plants and plants infected with 10 prevalent diseases (citrus greening, citrus canker, anthracnose, scab, greasy spot, melanose, sooty mold, nitrogen deficiency, magnesium deficiency, and iron deficiency) were used as materials. Initially, three datasets with white, natural, and mixed backgrounds were constructed to analyze their effects on the training accuracy, test generalization ability, and classification balance. This diversification of data significantly improved the model's adaptability to natural settings. Subsequently, by leveraging agricultural domain knowledge, a structured citrus disease features glossary was developed to enhance the efficiency of data preparation and the credibility of identification results. To address the underutilization of multimodal data in existing models, this study explored semantic embedding methods for disease images and structured descriptive texts. Convolutional networks with different depths (VGG16, ResNet50, MobileNetV2, and ShuffleNetV2) were used to extract the visual features of leaves. Concurrently, TextCNN and fastText were used to extract textual features and semantic relationships. By integrating the complementary nature of the image and text information, a joint learning model for citrus disease features was achieved. ShuffleNetV2 + TextCNN, the optimal multimodal model, achieved a classification accuracy of 98.33% on the mixed dataset, which represented improvements of 9.78% and 21.11% over the single-image and single-text models, respectively. This model also exhibited faster convergence, superior classification balance, and enhanced generalization capability, compared with the other methods. The image-text multimodal feature fusion network proposed in this study, which integrates text and image features with domain knowledge, can identify and classify citrus diseases in scenarios with limited samples and multiple background noise. The proposed model provides a more reliable decision-making basis for the precise application of biological and chemical control strategies for citrus production.

HSCCC Separation of Three Main Compounds from the Crude Extract of Dracocephalum Tanguticum by Using Dimethyl Sulfoxide as Cosolvent.

Separation of natural compounds directly from the crude extract is a challenging work for traditional column chromatography. In the present study, an efficient method for separation of three main compounds from the crude extract of Dracocephalum tanguticum has been successfully established by high-speed counter-current chromatography (HSCCC). The crude extract was directly introduced into HSCCC by using dimethyl sulfoxide as cosolvent. Ethyl acetate/n-butyl alcohol/0.3% glacial acetic acid (4: 1: 5, v/v) system was used and three target compounds with purity higher than 80% were obtained. Preparative HPLC was used for further purification and three target compounds with purity higher than 98% were obtained. The compounds were identified as chlorogenic acid, pedaliin and pedaliin-6″-acetate.

Separation of three anthraquinone glycosides including two isomers by preparative high-performance liquid chromatography and high-speed countercurrent chromatography from Rheum tanguticum Maxim. ex Balf.

Anthraquinone glycosides, such as chrysophanol 1-O-ß-d-glucoside, chrysophanol 8-O-ß-d-glucoside, and physion 8-O-ß-d-glucoside, are the accepted important active components of Rheum tanguticum Maxim. ex Balf. due to their pharmacological properties: antifungal, antimicrobial, cytotoxic, and antioxidant activities. However, an effective method for the separation of the above-mentioned anthraquinone glycosides from this herb is not currently available. Especially, greater difficulty existed in the separation of the two isomers chrysophanol 1-O-ß-d-glucoside and chrysophanol 8-O-ß-d-glucoside. This study demonstrated an efficient strategy based on preparative high-performance liquid chromatography and high-speed countercurrent chromatography for the separation of the above-mentioned anthraquinone glycosides from Rheum tanguticum Maxim.ex Balf.

An Efficient Protocol for Preparation of Gallic Acid from Terminalia bellirica (Gaertn.) Roxb by Combination of Macroporous Resin and Preparative High-Performance Liquid Chromatography.

In this article, macroporous resin column chromatography and preparative high-performance liquid chromatography were applied for preparation of gallic acid from Terminalia bellirica (Gaertn.) Roxb. In the first step, six kinds of resins were investigated by adsorption and desorption tests and AB-8 macroporous resin was selected for the enrichment of gallic acid. As a result, 20 g of gallic acid at a purity of 71% could be separated from 100 g of crude extract in which the content of gallic acid was 16.7% and the recovery of gallic acid reached 85.0%. In the second step, preparative high-performance liquid chromatography was selected to purify gallic acid. As a result, 640 mg of gallic acid at a purity of 99.1% was obtained from 1 g of sample in 35 min. The results demonstrated that macroporous resin coupled with preparative high-performance liquid chromatography was suitable for preparation of gallic acid from T. bellirica (Gaertn.) Roxb.

Application of high-speed counter-current chromatography combined with macroporous resin for rapid enrichment and separation of three anthraquinone glycosides and one stilbene glycoside from Rheum tanguticum.

In this paper, an efficient method was successfully established by the combination of macroporous resin (MR) and high-speed counter-current chromatography (HSCCC) for rapid enrichment and separation of aloe-emodin 8-O-ß-D-glucoside, emodin 1-O-ß-D-glucoside, emodin 8-O-ß-D-glucoside and piceatannol 4'-O-ß-D-(6″-O-gallate)-glucoside. Six kinds of macroporous resins were investigated in the first step and X-5 macroporous resin was selected for the enrichment of the target compounds. The recoveries of the target compounds reached 89.0, 85.9, 82.3 and 84.9% respectively after 40% ethanol elution. In the second step, the target compounds were separated by HSCCC with a two-phase solvent system composed of chloroform/ethyl acetate/methanol/water (8:1:6:5, v/v). The established method will be helpful for further characterization and utilization of Rheum tanguticum. The results demonstrate that MR coupled with HSCCC is a powerful technique for separation of bioactive compounds from natural products.

Application of an efficient strategy based on liquid-liquid extraction, high-speed counter-current chromatography, and preparative HPLC for the rapid enrichment, separation, and purification of four anthraquinones from Rheum tanguticum.

This study presents an efficient strategy based on liquid-liquid extraction, high-speed counter-current chromatography, and preparative HPLC for the rapid enrichment, separation, and purification of four anthraquinones from Rheum tanguticum. A new solvent system composed of petroleum ether/ethyl acetate/water (4:2:1, v/v/v) was developed for the liquid-liquid extraction of the crude extract from R. tanguticum. As a result, emodin, aloe-emodin, physcion, and chrysophanol were greatly enriched in the organic layer. In addition, an efficient method was successfully established to separate and purify the above anthraquinones by high-speed counter-current chromatography and preparative HPLC. This study supplies a new alternative method for the rapid enrichment, separation, and purification of emodin, aloe-emodin, physcione, and chrysophanol.