In China, low degree of automation seriously affects the working efficiency and quality in vegetable transplanting. As one of the most important vegetables in China even in the world, tomato was taken as the research object in this study. An automatic single-row transplanting device was designed, based on the statistical analysis of the physical and mechanical properties of tomato seedlings of a typical variety. Based on the technology of mechatronics, the device integrated the functions of transporting seedling tray, automatic seedling extraction and mechanical planting. The kinematics orthogonality solution combined with the dynamic sequence solution method was used to optimize and analyze the kinematic parameters of the automatic seeding mechanism, and the "sickle" trajectory was obtained. According to the position and movement requirement for taking and dropping seedling, the mechanical conditions and the working parameters of key execution parts were obtained by using analytic drawing method to analyze the mechanical condition of seedling collecting mechanism. The transplanting experiment was conducted at room temperature of 25°C, and the age and moisture content of the seedlings were 40 d and 55%, respectively. The results showed that the highest success rate was 92.59%, and the lowest rate of leakage was 23.13%, when the transplanting frequency was 60 plants/min. The lowest success rate was 77.78%, and the highest rate of leakage was 38.75%, when transplanting frequency was 120 plants/min. When the transplanting frequency is between 60-90 plants/min, the device can meet the requirement of high speed transplanting for potted vegetable seedling.

Interactions among users on social network platforms are usually positive, constructive and insightful. However, sometimes people also get exposed to objectionable content such as hate speech, bullying, and verbal abuse etc. Most social platforms have explicit policy against hate speech because it creates an environment of intimidation and exclusion, and in some cases may promote real-world violence. As users' interactions on today's social networks involve multiple modalities, such as texts, images and videos, in this paper we explore the challenge of automatically identifying hate speech with deep multimodal technologies, extending previous research which mostly focuses on the text signal alone. We present a number of fusion approaches to integrate text and photo signals. We show that augmenting text with image embedding information immediately leads to a boost in performance, while applying additional attention fusion methods brings further improvement.

No abstract

No abstract

No abstract