locate_object
Locate a named object in an image and return its coordinates. If absent, returns count=0. Supports pixel or normalized units and optional two-stage refinement for better precision.
Instructions
在图片中定位目标对象,返回坐标 primitives(让 LLM 输出坐标)。找不到会返回 count=0。
Tool call template
Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.
{
"id": 1,
"jsonrpc": "2.0",
"method": "tools/call",
"params": {
"arguments": {
"image": "<image>",
"target": "<target>"
},
"name": "locate_object"
}
}Input Schema
| Name | Required | Description | Default |
|---|---|---|---|
| image | Yes | ||
| coords | No | 返回坐标单位:pixel(默认,像素)或 norm(0-1000 归一化) | |
| refine | No | 两阶段精修:粗定位后裁切放大二次定位(更准,代价是每个候选多一次视觉调用),默认 false | |
| target | Yes | 要定位的目标,如:蓝色提交按钮 / 红色圆形 / 报错文字 |