Image artificial intelligence (AI) has recently become integral to the design process, and its potential for extensive use in the idea-image conversion stage by swiftly generating creative images has been demonstrated. Converting sketches into images during the design process is a positive step in leveraging AI image generators. This study explores factors of generating image prompts from sketches during the conversion process using AI image generators. We investigated the features of the AI generators and compared the image extraction features at different sketch stages using Midjourney and DALL-E 2 to find a workflow. Moreover, through image generation experiments, we identified effective ways to control the resulting images in sketch stages, such as sketch coloring and text representation. The findings of this study can offer guidance for the rapid transition from the abstract idea stage to the concrete digital image stage in the landscape design process. This will also help designers to considerably reduce the time and effort required to realize ideas and explore creative variations.